ww.sciencedirect.com

wat e r r e s e a r c h 7 5 ( 2 0 1 5 ) 1 6 4e1 8 7

Available online at w

ScienceDirect

journal homepage: www.elsevier .com/locate /watres

Review

Emerging desalination technologies for watertreatment: A critical review

Arun Subramani a,b,*, Joseph G. Jacangelo a,b

a MWH, 300 North Lake Avenue, Suite 400, Pasadena, CA 91101, USAb The Johns Hopkins University Bloomberg School of Public Health, Baltimore, MD 21205, USA

a r t i c l e i n f o

Article history:

Received 28 November 2014

Received in revised form

12 February 2015

Accepted 16 February 2015

Available online 26 February 2015

Keywords:

Nanotechnology membranes

Industrial water desalination

Energy minimization

Produced water treatment

Reverse osmosis

* Corresponding author. Chesapeake Energy2875.

E-mail address: arun.subramani@chk.comhttp://dx.doi.org/10.1016/j.watres.2015.02.0320043-1354/© 2015 Elsevier Ltd. All rights rese

a b s t r a c t

In this paper, a review of emerging desalination technologies is presented. Several tech-

nologies for desalination of municipal and industrial wastewater have been proposed and

evaluated, but only certain technologies have been commercialized or are close to

commercialization. This review consists of membrane-based, thermal-based and alter-

native technologies. Membranes based on incorporation of nanoparticles, carbon nano-

tubes or graphene-based ones show promise as innovative desalination technologies with

superior performance in terms of water permeability and salt rejection. However, only

nanocomposite membranes have been commercialized while others are still under

fundamental developmental stages. Among the thermal-based technologies, membrane

distillation and adsorption desalination show the most promise for enhanced performance

with the availability of a waste heat source. Several alternative technologies have also been

developed recently; those based on capacitive deionization have shown considerable im-

provements in their salt removal capacity and feed water recovery. In the same category,

microbial desalination cells have been shown to desalinate high salinity water without any

external energy source, but to date, scale up of the process has not been methodically

evaluated. In this paper, advantages and drawbacks of each technology is discussed along

with a comparison of performance, water quality and energy consumption.

© 2015 Elsevier Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165

2. Membrane-based technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166

2.1. Novel membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166

2.1.1. Nanocomposite membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166

2.1.2. Aquaporin membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169

2.1.3. Nanotube membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170

2.1.4. Graphene-based membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170

Corporation, 6100 North Western Avenue, Oklahoma City, OK 73118, USA. Tel.: þ1 405 935

(A. Subramani).

rved.

wat e r r e s e a r c h 7 5 ( 2 0 1 5 ) 1 6 4e1 8 7 165

2.2. Membrane-based processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171

2.2.1. Semi-batch RO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171

2.2.2. Forward osmosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172

3. Thermal-based technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173

3.1. Membrane distillation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173

3.2. Humidificationedehumidification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176

3.3. Adsorption desalination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177

3.4. Pervaporation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177

4. Alternative technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178

4.1. Microbial desalination cell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178

4.2. Capacitive deionization technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180

4.2.1. Membrane-based systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180

4.2.2. Flow through systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180

4.2.3. Hybrid systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180

4.2.4. Entropy battery systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180

4.3. Ion concentration polarization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181

4.4. Clathrate hydrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181

5. Application potential of emerging technologies in the treatment of challenging wastewater sources . . . . . . . . . . . . . . . 182

6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183

Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183

1. Introduction

Freshwater is a renewable resource, but increasing population

growth and population density has strained the ability of

many local supplies to sustainwater quantity requirements at

suitable levels of water quality. In response to the United

Nations prediction that 2e7 billion people will face water

scarcity by themiddle of the century (Hameeteman, 2013), the

water industry has become increasingly reliant upon desali-

nation of ocean and brackish water supplies. Desalination

processes are broadly categorized as thermal or membrane-

based technologies (Greenlee et al., 2009). Although thermal

desalination has remained the primary technology of choice

in the Middle East, membrane processes, such as reverse

osmosis (RO), have rapidly developed since the 1960's (Loeb

and Sourirajan, 1963) and currently surpass thermal pro-

cesses in new plant installations (Greenlee et al., 2009). The

primary drawback with desalination is associated with costs

(Subramani et al., 2011); those associated with electricity for

seawater desalination using RO are 30% of the total cost of

desalinated water. Higher energy consumption also translates

to a corresponding increase in greenhouse gas (GHG) emis-

sions (Raluy et al., 2005). Thus, reducing energy consumption

is critical for lowering the cost of desalination and addressing

environmental concerns about GHG emissions from the

continued use of conventional fossil fuels as the primary en-

ergy source for seawater desalination plants.

During desalination with RO membranes, brackish water

or seawater is pressurized against a semi-permeable RO

membrane that allows water to pass through while rejecting

salt. In order to produce desalinated water, the osmotic

pressure of the feed water needs to be exceeded. The feed

water to the RO is pressurized using a high pressure feed

pump to supply the necessary pressure to force water

through the membrane to exceed the osmotic pressure and

overcome differential pressure losses through the system

(Stover, 2007). In seawater desalination applications, an en-

ergy recovery device (ERD) in combination with a booster

pump is used to recover the pressure from the concentrate

and reduce the required size of the high pressure pump

(Stover, 2007). In brackish water applications, ERDs are

seldom utilized due to the low total dissolved solids (TDS)

levels, while certain full-scale plants have installed turbo-

chargers or isobaric devices to act as interstage booster

pumps (Drak and Adato, 2014).

A theoretical minimum energy exists that is required to

exceed the osmotic pressure and produce desalinated water.

As the salinity of the feed water or as feed water recovery

increases, the minimum energy required for desalination

also increases. For example, the theoretical minimum energy

for seawater desalination with 35,000 mg/L of salt and a feed

water recovery of 50% is 1.06 kWh/m3 (Elimelech and Phillip,

2011). However, the actual energy consumption is larger for

full-scale plants. The energy required for desalination using

RO membranes is a function of the feed water recovery,

intrinsic membrane resistance (permeability), operational

flux, feed water salinity and temperature fluctuations, prod-

uct water quality requirements, and system configuration

(Subramani et al., 2011). The lowest energy consumption re-

ported for an RO system is 1.58 kWh/m3 at a feed water re-

covery of 42.5% and a flux of 10.2 L m�2 h�1 (Seacord et al.,

2006). In addition, pre- and post-treatment contributes to

additional energy requirements (Wilf and Bartels, 2005).

Typically, the total energy requirement for seawater desali-

nation using RO (including pre- and post-treatment) is on the

order of 3e6 kWh/m3 (Semiat, 2008; Subramani et al.,

2014a,b).

wat e r r e s e a r c h 7 5 ( 2 0 1 5 ) 1 6 4e1 8 7166

Established techniques for minimizing energy usage for

RO are classified according to enhanced system design, high

efficiency pumping and energy recovery (Subramani et al.,

2011). Desalination energy minimization techniques have

also focused on feedback controls linked to the feed water

salinity to optimize energy consumption in RO applications

(Bartman et al., 2010; Gao et al., 2014). Control schemes uti-

lized real-time updates on RO membrane permeability to

adjust feed pressure requirements to maintain operational

flux and lowest specific energy consumption (Gao et al.,

2014). In recent years, significant improvements in the salt

rejection capacity and permeability of membranes for treat-

ing high salinity feed waters have been achieved. Today, the

average energy consumption for seawater desalination using

RO is still higher than the theoretical minimum energy

required and improvements in desalination membranes

show promise for lowering energy consumption. New desa-

lination technologies offer reduced feed pressure re-

quirements while maintaining salt rejection (Subramani

et al., 2011; Pendergast and Hoek, 2011). In an analysis of

the impact of RO permeability and energy minimization, Zhu

et al. (2009) found that when the cost of energy and mem-

brane is considered, the benefit of developing membranes

with higher permeability is less likely to be a significant

driver for lower seawater desalination costs. Thus, other

factors such as process improvements, development of

fouling resistant membranes, lesser pretreatment and effec-

tive brine management need to be considered to lower

overall desalination costs (Zhu et al., 2009, 2010).

Three major reviews on emerging desalination technolo-

gies and methods to reduce energy consumption have been

recently published (Subramani et al., 2011; Pe~nata and

Garcıa-Rodrıguez, 2012; Elimelech and Phillip, 2011).

Subramani et al. (2011) performed a detailed review of energy

optimization techniques available for seawater desalination

along with a review of renewable energy resource utilization.

Pe~nata and Garcıa-Rodrıguez (2012) reviewed the current

trends and future prospects for seawater desalination to

improve process performance and obtain high productivity

but their review considered only RO technology and limited

information was provided on innovative technologies.

Elimelech and Phillip (2011) reviewed strategies to reduce

energy consumption for seawater desalination and

concluded that optimization of pretreatment and post treat-

ment is the best method for energy reduction. However,

pretreatment and post treatment processes contribute less

than 10% of the overall energy consumption during seawater

desalination (Wilf and Bartels, 2005). In the above reviews,

only certain innovative technologies, such as nanocomposite

reverse osmosis, carbon nanotubes and forward osmosis

membranes, were included. Numerous emerging technolo-

gies have been developed recently with more performance

data published in the literature specific to desalination. Thus,

in our review a comprehensive description of emerging

desalination technologies is provided to complement those

previously published.

Specifically, the objective of this review is to critically

compare emerging desalination technologies that show

promise for energy minimization and improved performance;

it is divided according to the principle of operation

(membrane-based and thermal-based) while discussing

alternative technologies as well. The first category consists of

new desalination technologies that utilize membranes for

separation and the second on those that utilize a temperature

gradient for separation to occur. Finally, alternative technol-

ogies are reviewed and consist of those that utilize mecha-

nisms different from membranes or thermal technologies.

Application potential of these three technology categories in

municipal and industrial water sectors are discussed along

with their implementation challenges.

2. Membrane-based technologies

In this section, emerging desalination technologies based on

membrane processes is discussed. This section is divided into

two sub-sections: the first sub-section deals with develop-

ment and application of new generation membrane material

for desalination and the second sub-section provides a review

of emerging technologies based on membranes for desalina-

tion. Certain membrane-based technologies are close to

commercialization whereas others are still under develop-

mental stages. A comparison of membrane-based technolo-

gies is provided in Table 1.

2.1. Novel membranes

2.1.1. Nanocomposite membranesA thin film nanotechnology (TFN) membrane incorporates

Linde type A zeolite nanoparticles into the thin film layer of

the membrane to enhance permeability of water while

maintaining salt rejection (Jeong et al., 2007). Linde type A is

an alumino silicate zeolite that exhibits a three dimensional

pore structure with the pores perpendicular to each other in

the x,y and z planes. The utilization of these nanoparticles

increases the flux through the membrane, and thus provides

an opportunity to reduce energy consumption through a

lowered feed pressure while maintaining the same water

production levels. TFN technology utilizes an interfacial

polymerization process with benign nanoparticles dispersed

in one ormore of themonomer solutions in order to create the

nanocomposite membranes (Kurth et al., 2014). These mem-

branes have a measurably smoother, more hydrophilic, and

more negatively charged surface in comparison with a typi-

cally pure polyamide thin film composite (TFC)membrane due

to the presence of the nanoparticle pores (Jeong et al., 2007).

These surface properties improve the permeability of the

membranes, essentially creating molecular, hydrophilic tun-

nels across the membrane matrix which water preferentially

passes through, while the greater negative charge on the

nanoparticle pore walls enhance ion exclusion, and thus

maintain salt rejection (Pendergast et al., 2013). The hydro-

philic nanoparticles make the membrane as a whole e which

is typically hydrophobic in its pure polyamide composition e

more hydrophilic, leading to less susceptibility to membrane

fouling. In an early study on TFN technology, at the highest

levels of nanoparticle loading, permeability was reported to be

almost twice that of a pure polyamide thin film composite

formed by the same polymerization process while the salt

rejection was maintained for the TFN membrane (Jeong et al.,

Table 1 e Comparison of membrane-based technologies.

Technology Advantages Drawbacks Recovery range Feed waterquality

Treated waterquality

Energyconsumption

Cost impacts

Nanocomposite

Membranes

- Increased permeability

while maintaining salt

rejection.

- Reduced feed pressure

requirement.

- Reduced footprint at high

flux operation.

- More expensive mem-

brane elements.

- Retrofitting of existing

plants may require a var-

iable speed drive (VFD) on

pumps.

Variable, 40e50%. 32,000

e34,000 mg/L

(TDS). Brackish

water

membranes are

under

development.

Similar to TFC RO

membranes.

1.73 kWh/m3

e2.49 kWh/m3

for seawater

desalination with

TDS of 32,000mg/

L (Subramani

et al., 2014a).

- Potential to reduce costs

through: decreased en-

ergy with same water

production, increased flux

with same feed pressure,

decreased footprint with

same water production.

- Currently applicable only

for seawater desalination.

Aquaporin

membranes

- High permeability (an

order of magnitude

higher than commercial

RO membranes).

� 100% rejection of solute

molecules.

- Osmotically-driven pro-

cess without the need for

applied pressure.

- Synthetic approach to

producing and purifying

aquaporin in large quan-

tities is essential.

- Limited experimental

data with real feed water

sources.

- Chemical resistance of

aquaporin is unknown.

Structural strength of

membranes is unknown.

Not known. No limitation on

feed water TDS.

100% rejection of

TDS.

Not known.

Expected to be

low due to

absence of feed

pressure

requirement.

Not known. Technology still

at bench-scale level. Costs

will depend on the

production of aquaporins

on a large scale to

synthesize membranes.

Nanotube

membranes

- High permeability (10

times higher than com-

mercial RO membranes);

- High rejection of salt.

- Packing density of nano-

tubes on substrate not

known for practical

applications.

- Limited experimental

data with real feed water

sources.

- Rejection of specific con-

taminants is not known

and functionalization of

nanotubes is not known.

- Stability of nanomaterial

within support/substrate

layer is not known.

- Associated health risks

with release of nano-

material into treated

water stream is not

known.

Not known. Not known. More than 95%

rejection of salt.

Not known.

Expected to be

similar to RO

membranes.

Desalination

applications

limited by

thermodynamic

restriction.

Not known. Technology still

at bench-scale level. Costs

will depend on packing

density of nanotubes.

Graphene-based

membranes

- Good mechanical

properties.

- Fast water transport and

high rejection capability.

- Requires applied

pressure.

- Only bench-scale and

modeling studies have

been performed.

Not known. Not known. Up to

seawater salinity

must be possible.

Not known. Not known.

Expected to be

similar to RO

membranes.

Desalination

Not known. Technology still

at bench-scale level. Costs

will depend on the ability to

synthesize large quantities

(continued on next page)

water

research

75

(2015)164e187

167

Table 1 e (continued )

Technology Advantages Drawbacks Recovery range Feed waterquality

Treated waterquality

Energyconsumption

Cost impacts

- Membranes can be tuned

to be ion selective.

applications

limited by

thermodynamic

restriction.

of graphene material to

form membranes.

Semi-batch RO - Reduced energy

consumption.

- Energy recovery devices

are not required.

- Higher feed water recov-

ery for brackish and

wastewater applications

(>90%).

- Higher operational flux

(>30%) with fewer mem-

brane elements.

- Reduced plant footprint.

- Less susceptibility to

fouling due to high cross-

flow velocity. Potential for

lesser pretreatment.

- Only pilot and demon-

stration scale testing data

is available.

- Effect of high cross flow

and brine circulation on

membrane life is

unknown.

- Up to 55% for seawater

desalination.

- Up to 95% for brackish water and

wastewater desalination.

Up to 45,000mg/L

(TDS).

Similar to

conventional RO.

2.74 kWh/m3 for

seawater

desalination with

TDS of 37,000mg/

L (Subramani

et al., 2014b).

- The utilization of fewer

membranes and no

requirement for ERDs re-

duces capital expendi-

tures substantially for

seawater desalination.

- Lower fouling potential

could reduce operating

costs with respect to fre-

quency of chemical

cleanings.

Forward

osmosis

- Osmotically-driven pro-

cess without the need for

applied pressure.

- Multiple applications.

- Applicable for high

salinity desalination.

- Potential savings in en-

ergy consumption when

combined with RO.

- Can utilize waste heat

source for regeneration of

draw solution.

- High feed water recovery.

- Lower fouling potential

due to lack of applied

pressure. Potential for

lesser pretreatment.

- Limited full-scale

installations.

- Difficult to choose optimal

osmotic agent (draw

solution).

- Lower flux rates than RO

leading to higher mem-

brane area requirement.

- Requires membranes

specific for FO

applications.

�91.9% at leachate treatment

plant.

�35% at full-scale seawater facility.

�96% in FO/RO hybrid systems for

wastewater.

500 to

175,000 mg/L

(TDS).

Similar to TFC RO

membranes.

Similar to RO for

seawater

desalination.

- The costs depend on the

application of the tech-

nology. In strictly FO ap-

plications with gaseous

mixtures as draw solu-

tion, the primary cost will

be thermal energy

required for recovery/re-

concentration of the draw

solution.

- In FO/RO hybrid systems

using reuse water as feed

to FO and RO for seawater

desalination, the cost of

desalinating water will be

lower as less electrical

energy will be required in

the RO stage due to the

dilution of feed seawater.

water

research

75

(2015)164e187

168

wat e r r e s e a r c h 7 5 ( 2 0 1 5 ) 1 6 4e1 8 7 169

2007). The highest TFN element permeate flow rate reported

was 52 m3/d and a minimum NaCl rejection of 99.7%

(NanoH2O, Inc., 2011).

At bench scale, Pendergast et al. (2013) evaluated two

different types of TFN membranes, which were developed to

improve water permeability and salt rejection along with

improved resistance to physical compaction. When zeolite

nanoparticles were embedded within the polyamide active

layer, compaction of the membrane was suppressed. Incor-

poration of the nanoparticles in the polysulfone support layer

also resulted in a smaller water contact angle and higher ul-

timate tensile strength when compared to an unmodified

polysulfone support layer. Although water permeability was

enhanced with the incorporation of zeolite nanoparticles in

the polyamide matrix while maintaining the salt rejection,

comparisons were made with hand cast TFC membranes. It is

important to note that results based on these bench-scale

studies may not be directly comparable to TFC membranes

obtained from commercial manufacturers. Thus, other

studies have compared TFN membranes with commercially

available TFC membranes.

Hofs et al. (2013) compared TFN and TFC membrane ele-

ments at pilot scale and reported that TFN elements exhibited

1.4 times higher permeability when compared to TFC mem-

brane elements. In this study, a TFN membrane element

SW365ES was compared with a TFC membrane element

SW30XHR-440i from DowFilmtec. Although the TFN mem-

brane exhibited higher water permeability while maintaining

similar salt rejection, boron rejection and low molecular

weight nitrosamines rejection was lower when compared to

the TFC membrane. New generation TFN membranes with

improved boron rejection have been recently introduced in

the market but the water permeability of these new genera-

tion TFN membranes are similar to TFC membranes. It is also

unclear if the concentration of nanoparticles (<6 wt%) in the

TFN membrane's active layer results in the increased

permeability.

In a recent study, Subramani et al. (2014a) compared

NanoH2O's TFN membranes Qfx400R and Qfx400ES with

DowFilmtec's TFC membranes SW30XLE and SW30ULE for

desalination of Pacific Ocean seawater with a TDS of

34,000 mg/L. The specific energy consumption for the TFN

membranes was 2.24e2.55 kWh/m3 for fluxes of

11.9e15.3 L m�2 h�1 and a system recovery of 40e55%. The

specific energy consumption for the TFC membranes was

2.28e2.61 kWh/m3 for the same flux and recovery conditions.

Thus, the savings in energy consumption was less than 6%

using TFN membranes. The use of TFN membranes also

resulted in lower boron rejection when compared to TFC

membranes. However, in seawater desalination a second pass

RO system with brackish water membranes operated at a pH

of approximately 10.3 is utilized to achieve very low levels

(<0.5 mg/L) of boron in the treated water stream. Selection of

TFN RO membranes over TFC RO membranes also requires a

careful consideration of life cycle costs. When the TFN RO

membrane element cost is higher, capital costs will be higher.

However, life cycle costs of the plant could potentially be

lower due to savings in energy costs over the plant operational

life.

2.1.2. Aquaporin membranesAquaporins are the protein channels that control water flux

across biological membranes. They are found widely in

human tissues with the purpose of rapid, passive transport of

water molecules across cell membranes (Pendergast and

Hoek, 2011). Water movement in an aquaporin is mediated

by selective, rapid diffusion and caused by osmotic gradients

(Agre, 2003). Aquaporin-1 (AQP1), with selective extracellular

and intracellular vestibules at each end, allows water mole-

cules to pass rapidly in a single-file line, while excluding

proteins and ions by an electrostatic tuningmechanism (Agre,

2003; Sui et al., 2001). The result leads to only water molecules

being transported through the aquaporin channels and

charged ions being rejected (Sui et al., 2001; Bowen, 2006).

Aquaporin membranes are considered to be 100 times

more permeable than commercial RO membranes (Kaufman

et al., 2010). Highly permeable and selective membranes

based on the incorporation of the functional water channel

protein Aquaporin Z into a novel triblock copolymer have

been shown to have significantly higher water transport than

existing RO membranes (Kumar et al., 2007). Kumar et al.

(2007) utilized Aquaporin-Z from Escherichia coli bacterial

cells in a polymeric membrane. The aquaporin was selected

based on the ability for high water permeation and selectivity.

A symmetric triblock copolymer with a high hydrophobic to

hydrophilic block ratio was selected to mimic a lipid-bilayer

membrane. The resulting protein-polymer membrane

demonstrated over an order of magnitude increase in water

permeability over a purely polymeric membrane. The aqua-

porin membrane also rejected glucose, glycerol, salt and urea

to detection limits. In another study, Zhu et al. (2004) simu-

lated the permeation of water molecules through AQP1 (Zhu

et al., 2004). Two factors were proposed for the transport of

water molecules: molecular and diffusion permeability. The

former was a result of concentration differences leading to

mass transfer while the latter was due to random movement

ofmolecules with no net transfer (Zhu et al., 2004). Wang et al.

(2012) constructed an aquaporin membrane on a porous

support by a combination of pressure assisted vesicle

adsorption and covalent-conjugation-driven vesicle fusion.

The water flux through the membrane ranged between 34.1

and 73.8 L m�2 h�1. The experimental obtained water flux was

<10% of the theoretical water flux obtained using computer

modeling.

Studies have also been performed where aquaporins were

deposited onto commercially available membranes. Kaufman

et al. (2010) deposited an aquaporin onto commercially NF270

and NTR7450 membranes via vesicle fusion at a pH of 2.0. The

investigators demonstrated supported lipid bilayers formed

atop dense water permeable nanofiltration (NF) membranes

that can be operated under a mechanical driving force as with

RO membranes. NF membranes were chosen as the support

due to their high permeability and low surface roughness that

allowed for minimal distortion of the lipid bilayer.

Membranes based on aquaporins show promise for desa-

lination where the driving mechanism is an osmotic pressure

gradient (salt concentration), rather than a mechanically

applied pressure gradient as in RO (Pendergast and Hoek,

2011). With 75% coverage of aquaporins on a membrane, an

wat e r r e s e a r c h 7 5 ( 2 0 1 5 ) 1 6 4e1 8 7170

order of magnitude increase in hydraulic permeability has

been predicted (Kaufman et al., 2010). Due to the absence of

applied pressure, energy consumption is expected to be sub-

stantially lower as compared to RO membranes. Due to the

difficulty of attaining large quantities of proteins and pro-

ducing large areas of membrane material, aquaporin-based

membranes are not widely available for commercialization

(Pendergast and Hoek, 2011).

2.1.3. Nanotube membranesCarbon nanotubes have been evaluated for desalination due

to their rapid water transport properties, large surface area

and ease of functionalization (Majumder et al., 2005; Hummer

et al., 2001; Humplik et al., 2011; Hinds et al., 2004; Corry, 2008).

Desalination energy consumption using carbon nanotubes

can be significantly lower when compared to RO due to the

nanotubes' water mass transport being 2 to 5 times higher

than theoretical predictions by the HagenePoiseuille equation

(Holt et al., 2006; Ahadian and Kawazoe, 2009). Water and ions

are transported through membranes formed from carbon

nanotubes that range in diameter from 6 to 11 A. The high flow

rate has been attributed to the atomic smoothness and mo-

lecular ordering of the nanotubes through which water mol-

ecules are passed on a one dimensional single-file procession

(Hummer et al., 2001; Kalra et al., 2003). The challengewith the

use of carbon nanotubes for desalination has been due to the

complexities involved in the fabrication of sub-nanometer

tubes.

Carbon nanotubes have been incorporated onto several

types of substrates using catalytic chemical vapor deposition

(CVD) (Holt et al., 2006; Kim et al., 2007; Yoshikawa et al., 2008).

Results for desalination applications showed that narrow

carbon nanotubes could completely reject ions due to the

large energy barrier at the nanotube openings created by

stable hydrogen bond formation (Corry, 2008). On the con-

trary, water does not form stable hydrogen bonds with the

nanotubes and permeates rapidly. Membranes incorporating

carbon nanotubes have been found to be promising candi-

dates for water desalination, as the size and uniformity of the

tubes can achieve the desired salt rejection (Corry, 2008). A 10-

fold permeability increase is expected using a carbon nano-

tube RO membrane (Corry, 2008; Holt et al., 2006; Sholl and

Johnson, 2006).

Ion rejection by carbon nanotube membranes is governed

by the steric effects between nanopores and the hydrated

diameter of ions along with the Donnan equilibrium of the

membrane surface (Ahn et al., 2012). Corry (2008) found that

the rejection of ions by carbon nanotubes was dependent on

the pore size of the nanotube (Corry, 2008). As the inner

diameter of the nanotubes increased from 0.32 nm to 0.75 nm,

salt rejection efficiencies of the membrane declined from

100% to 58%. However, these results were based on molecular

dynamics simulation and not based on actual experimental

results. Ahn et al. (2012) found that the rejection efficiency of a

carbon nanotube membrane improved when the surface

charge of themembranewas elevated to increase electrostatic

interactions. Thus, modifying the surface properties of the

carbon nanotube could result in higher desalination effi-

ciency. Compared to conventional membranes, another

advantage is the longer lifetimes due to excellent mechanical

properties (Salvetat et al., 1999a,b). Hilder et al. (2009) showed

that boron nitride nanotubes had superior water flow prop-

erties when compared to carbon nanotubes while also

achieving 100% salt rejection (Hilder et al., 2009). The use of a

nanotube radius of 4.14 A can functionalize the membrane to

become cation-selective. When a nanotube radius of 5.52 A

was used, the membrane was functionalized to become

anion-selective (Hilder et al., 2009).

Carbon nanotubes have also been evaluated for their salt

adsorption capacity. Yang et al. (2013) showed that plasma

treatment of carbon nanotubes resulted in ultrahigh salt

adsorption capacity exceeding 400% by weight. Modified car-

bon nanotubes were fabricated with the deposition of a thin

layer of nanotubes onto a mixed cellulose ester porous sup-

port (Yang et al., 2013). The adsorption capacity of these

modified nanotubes was two orders of magnitude higher than

activated carbon material. The salt adsorption capacity was

recovered completely by a tap water rinse. The increase in salt

adsorption capacity of the modified nanotubes was attributed

to the defective sites created on the surface due to the plasma

treatment. In addition to the high surface area, the modified

surface enhanced surface hydrophilicity and ion binding

properties due to the functionalization of carboxylic and hy-

droxyl groups (Yang et al., 2013). Since the salt is adsorbed

rather than rejected, there is no requirement for applied

pressure. Hence, the energy consumption can be reduced

substantially.

Significant advantages of aligned nanotube membranes

over conventional membranes through reduced hydraulic

driving pressure, and lower energy costs have been reported,

but the productivity is limited by osmotic pressure via ther-

modynamic restriction (Song et al., 2003). It is also uncertain if

the nanotubes can be aligned with a high packing density to

obtain the projected permeability. Carbon nanotubes are a

material that is producible in large quantities; however,

fabrication of large surface areas after incorporation of

nanotubes will be a key step to enabling their commerciali-

zation (Pendergast and Hoek, 2011).

2.1.4. Graphene-based membranesGraphene-based membranes are being developed for desali-

nation due to their fast water transport properties and good

mechanical properties (Nair et al., 2012; Xue et al., 2013; Choi

et al., 2013; Mi, 2014). Similar to the water permeation mech-

anism in carbon nanotubes, two-dimensional graphene

nanocapillaries allow a low-friction flow of monolayer water

with size exclusion as the dominant sieving mechanism (Nair

et al., 2012). Nair et al. (2012) prepared graphene oxide (GO)

sheets by dispersing graphene through sonication and for-

mation of laminates by spray-coating or spin-coating (Nair

et al., 2012). Although graphene is impermeable to water

molecules, transport occurs through capillaries and can be as

fast as water molecule transport through an open aperture.

The capillaries formed within graphene laminates are attrib-

uted to the functional groups, such as hydroxyl and epoxy,

which are responsible for the creation of nanocapillaries (Nair

et al., 2012). Such groups create a cluster and leave large,

percolating regions of graphene sheets not oxidized. These GO

laminates have spaces formed between non-oxidized regions

of graphene sheets. Joshi et al. (2014) found that GO sheets are

wat e r r e s e a r c h 7 5 ( 2 0 1 5 ) 1 6 4e1 8 7 171

vacuum-tight in the dry state but swell and act as molecular

sieves when immersed in water and reject all solutes with a

hydrated radii larger than 4.5 A. Nanocapillaries are opened

up in the hydrated state and result in a low frictional flow of

water (Joshi et al., 2014).

Xue et al. (2013) studied graphyne, a one atom thick carbon

allotrope of graphene. Graphyne was formed by replacing

certain carbonecarbon bonds in graphene with acetylenic

linkages to form a-graphyne, b-graphyne, g-graphyne and its

analogs. Results proved that 100% rejection of common ions

present in seawater were rejected by the graphyne mono-

layers with a water permeability that was two orders of

magnitude higher than commercially available RO mem-

branes (Xue et al., 2013). Molecular dynamics and computer

simulations have been used for studying the transport of

water molecules through GO layers and it's application in

desalination. Nicolai et al. (2014) utilized molecular dynamics

and computer simulations to study the transport of water

molecules through GO layers. They showed that water

permeability through GO frameworkmembranes increases by

two orders of magnitude when the capillary pore size in-

creases with 100% salt rejection capability.

O'Hern et al. (2014) showed that the selectivity of GO

membranes can be tuned through the generation of sub-

nanometer pores. Oxidative etching resulted in pores with

diameters of 0.40± 0.24 nmand densities exceeding 1012 cm�2.

At short oxidation times, the pores were cation selective with

electrostatic repulsion due to the negatively charged func-

tional groups in the pore edges. At longer oxidation times, the

pores prevented the transport of larger organic molecules but

allowed the transport of salt, indicating steric size exclusion

(O'Hern et al., 2014). In an effort to fabricate GOmembranes for

desalination applications, Choi et al. (2013) created an as-

sembly of GO nanosheets on polyamide (PA) thin film com-

posite (TFC) ROmembranes. The GOmaterial was coated onto

the PA-TFCmembrane surface via layer-by-layer deposition of

oppositely charged nanosheets which resulted in increased

hydrophilicity and reduced surface roughness of the RO

membrane (Choi et al., 2013). The altered properties of the RO

membrane due to the GO nanosheets resulted in enhanced

resistance to protein fouling and increased chlorine resistance

(Choi et al., 2013; Kim et al., 2013a,b,c).

Graphene can be mass produced and it is possible to make

laminates with high mechanical strength and flexibility (Nair

et al., 2012). Functionalized graphene has also been utilized as

forward osmosis membranes to decrease internal concentra-

tion polarization (ICP) and increase water flux (Gai and Gong,

2014; Gai et al., 2014). However, water productivity, similar

to carbon nanotubes, will be limited by osmotic pressure via

thermodynamic restriction. Commercialization of graphene-

based membranes for desalination applications will depend

on the ability to synthesize large quantities of graphene ma-

terial and the mechanical strength of the nanolayers when an

applied hydrostatic pressure is present.

2.2. Membrane-based processes

2.2.1. Semi-batch ROThe semi-batch RO process combines raw feed water with

circulating concentrate at defined ratios (Song et al., 2012).

The process utilizes a combination of variable operating

pressure with internal concentrate recirculation and a

membrane configuration consisting of only three or four el-

ements per pressure vessel to reduce energy consumption

(Efraty et al., 2011). This process lowers the feed pressure

required for desalination, and hence, reduces energy con-

sumption. The process starts with a low initial feed pressure

that is gradually increased, while equal pressurized feed and

permeate flow rates aremaintained. In a typical 6.5 min cycle,

the operating pressure can vary between 40 and 70 bar (Stover

and Efraty, 2011). The recirculation of concentrate allows a

feed water recovery of 50% or more for seawater desalination

and more than 90% for brackish water desalination. The feed

water recovery depends on the time of concentrate recircu-

lation. Next to the re-circulation loop is an isobaric chamber

filled with feed water. At the end of a cycle, valves open and

the membranes are flushed with the pressurized water from

the isobaric chamber without stopping the pumps or

permeate flow. The side chamber is then closed off, depres-

surized by letting out a few drops of concentrate, recharged

with feed water, and re-pressurized with water from the

desalination loop. No pressurized concentrate stream is

released from the process. Thus, the pressure available in the

concentrate is (almost completely) utilized before discharge

and energy recovery devices (ERDs) are not required (Stover

and Efraty, 2011). As the system is operated at high cross-

flow velocities, fouling of the membranes is potentially

reduced.

The semi-batch RO process has been utilized for both

brackish and seawater desalination. Song et al. (2012)

demonstrated that the semi-batch RO process consumed

similar energy when compared to full-scale brackish water

desalination plants. For brackish water desalination, a SEC of

1.0 kWh/m3 was reported using the semi-batch RO system at a

flux of 26.1 L m�2 h�1. However, at a flux of 43.7 L m�2 h�1, the

SEC of semi-batch RO system was 150% of a full-scale RO

system operated with a conventional configuration (Song

et al., 2012). Stover (2013) utilized semi-batch RO for brackish

water desalination and achieved a SEC of 0.64e0.76 kWh/m3.

For the semi-batch RO system, only four elements were uti-

lized per pressure vessel when compared to seven elements

per pressure vessel for the conventional RO configuration. The

semi-batch RO process was operated an average flux of

20.1e28.4 L m�2 h�1 and a final feed water recovery of 94%

(Stover, 2013).

The semi-batch RO process has been commercialized as

closed-circuit desalination (CCD) by Desalitech, LLC and re-

ported to reduce desalination energy consumption up to 20%

(). Subramani et al. (2014b) evaluated the semi-batch RO pro-

cess for seawater desalination and compared its specific en-

ergy consumption with a conventional RO system for the

same operating conditions using TFN RO elements. The semi-

batch RO process with new TFN RO membranes exhibited

12e16% savings in SEC when compared to a conventional RO

process for seawater desalination. For an operational flux of

12.7 L m�2 h�1 and a feed water recovery of 44% and 53%, the

SEC for the semi-batch system was 2.16 kWh/m3 and

2.24 kWh/m3, respectively at a feed water temperature be-

tween 16 and 18 �C. For the same operational flux and feed

water recovery of 44% and 53%, the theoretical SEC for a

wat e r r e s e a r c h 7 5 ( 2 0 1 5 ) 1 6 4e1 8 7172

conventional RO system (based on modeling) was 2.46 kWh/

m3 and 2.67 kWh/m3, respectively (Subramani et al., 2014b).

The permeate water quality of the semi-batch RO system

was also similar to a conventional RO system. Another

consideration during the selection of a semi-batch RO is the

equipment requirement (Subramani et al., 2014b). The semi-

batch RO system requires additional pressure vessels that

are used as side conduits to replace the systemwith fresh feed

water and reject brine to enable the desalination cycle to be

continuous. Although an ERD is not required for the semi-

batch RO system, additional equipment in the form of auto-

mated pneumatic valves are required to continuously operate

the system. The cost of these additional valves must be

considered during evaluation of the capital cost. All the

pumps utilized in the conventional RO system are also

required for the semi-batch RO system with the addition of a

side conduit pump.

2.2.2. Forward osmosisForward osmosis (FO), a process that has been recognized as a

practical commercial process since the 1930's (Semiat, 2008)

and has been investigated for a wide range of applications to

desalinate water. In the FO process, instead of using hydraulic

pressure, as in conventional RO desalination processes, a

concentrated draw solution is used to generate high osmotic

pressure, which pulls the water across a semipermeable

membrane from the feed solution (McCutcheon et al., 2005).

The draw solutes are then separated from the diluted draw

solution to recycle the solutes and to produce a final product

water. Amixture of ammonia and carbon dioxide gas has been

used as the predominant draw solution (McCutcheon et al.,

2006). When ammonia and carbon dioxide are mixed in the

right proportion, a solution with a high osmotic pressure can

be formed. This solution has been used for drawing water

saline feeds. The advantage of using such a mixture is that it

has been shown to have the ability to be regenerated when

heated, and reused for the FO process. Thus, the FO can be

considered as a combination of membrane and thermal

processes.

While a variety of osmotic agents can be utilized, it is

imperative that draw solutions are non-toxic, stable, near

neutral pH, highly soluble to avoid precipitation, and can be

removed from water at a low cost using existing technology

(Cath et al., 2006; Liu et al., 2010; Achilli et al., 2010; Yen et al.,

2010; Klaysom et al., 2013). Volatile solutes such as KNO3, SO2,

or NH3/CO2 mixtures are viable osmotic agents as their tem-

perature dependent solubility allows for their thermal sepa-

ration from the draw solution and subsequent reuse.

Magnetoferritin is also a viable osmotic agent for reuse as this

material can easily be separated from draw solutions with the

application of a magnetic field (Liu et al., 2010). However,

while reusing these osmotic agents minimizes waste, the

separation of these agents from draw solutions remains the

primary consumer of energy in the FO process (Zhao et al.,

2012).

To further decrease FO energy requirements, draw solu-

tions that do not require separation treatment have been

developed.When fertilizers (i.e. KCl, NaNO3, Ca(NO3)2, etc.) are

employed as the osmotic agent, the product of FO desalination

is a diluted draw solution that can be applied to crops via

fertigation. As chemical fertilizers continue to be widely used,

this method serves as an effective and cost-efficient way to

supply water and nutrients to crops (Phuntsho et al., 2011). In

a similar FO application, sugars (glucose, fructose, sucrose)

and partially dehydrated food have been used as osmotic

agents in portable applicationswhere the drawwater does not

require treatment and becomes a nutrient solution for end use

(Liu et al., 2010). In another approach, switchable polarity

solvents that can be mechanically separated have been eval-

uated. Stone et al. (2013) utilized a mixture of CO2, water and

tertiary amines as draw solutes in forward osmosis. The pri-

mary advantage with this method lies in the utilization of

waste CO2 at atmospheric pressure to alter the properties of

switchable polarity solvents (Jessop et al., 2005, 2012). Once

the draw solution is diluted, the switchable polarity solvent is

mechanically separated using simple low-pressure filtration

techniques. In this approach, 1 atm of CO2 along with mild

heating results in the conversion of the switchable polarity

solvents from polar to nonpolar phase, which is then followed

by mechanical separation. The energy consumption with

switchable polarity solvents has been reported to be 35e48%

lower than using NH3/CO2 for the draw solution (Wilson et al.,

2013). When membranes with cellulose triacetate (CTA) were

used, the draw solution containing switchable polarity sol-

vents degraded the membrane. However, polyamide mem-

branes were stable and did not degrade (Stone et al., 2013).

Thus, membrane material compatibility needs to be further

studied before this process is applied at larger scales.

In certain FO applications, saline water is used as the draw

solution rather than the feed solution. The most simple of

these applications is FO for RO pretreatment. In this case,

seawater serves as the draw solution and freshwater (or a

solution with lower osmotic pressure) serves as a feed solu-

tion which will pressurize and dilute seawater for more

favorable RO conditions (Cath et al., 2009). This pretreatment

process greatly reduces the energy required to desalinate

water. In a similar process, pretreatment of RO water can be

carried out using impaired water (i.e. wastewater, leachate,

etc.) as the feed solution (Cath et al., 2009). The benefit of using

this type of water is that in addition to diluting RO feed water

to more favorable operating conditions, impaired water is

concentrated formore effective handling. In a similarmanner,

a novel process that uses ocean water to dewater an algae/

nutrient solution for the generation of algal biofuels has been

under investigation (Hoover et al., 2011).

In addition to choosing the optimal osmotic agent, there

are a variety of membrane types and configurations that will

affect FO performance. As internal and external concentration

polarization (ICP/EXP) can significantly affect performance,

ideal FO membranes should be designed with a high density

active layer that will achieve an elevated solute rejection, a

thin support layer with minimal porosity that will limit ICP,

and hydrophilic properties to maximize flux and minimize

fouling (Coday et al., 2013). Examples of membranes used for

FO include cellulose triacetate (CTA) with embedded polyester

mesh, aromatic polyamide polymer RO membranes, and

proprietary membranes developed to minimize EXP/ICP/

fouling and maximize flux and solute rejection (Cath et al.,

2006; Miller and Evans, 2006). In order to improve perfor-

mance of the FO process, membranes have also been

wat e r r e s e a r c h 7 5 ( 2 0 1 5 ) 1 6 4e1 8 7 173

fabricated in hollow fiber and flat sheet configurations. Fang

et al. (2012) fabricated hollow fiber FO membranes with poly

(amideeimide) material and created a double-skinned fiber

with an inner skin and outer skin similar to an RO and NF

membrane, respectively. A 2.0 M NaCl solution was used as

the draw solution. The hollowfibermembrane exhibited a flux

of 41.3 L m�2 h�1 by minimizing internal concentration po-

larization (Fang et al., 2012). Qui et al. (2012) fabricated flat

sheet FO membranes with polyamideeimide via phase

inversion followed by polyelectrolyte polyethyleneimine post-

treatment to form nanofiltration-like rejection layer with a

positive charge. These flat sheetmembranes achieved a water

flux of 29.7 L m�2 h�1. When the rejection layer was oriented

towards the draw solution, a water flux of 54 L m�2 h�1has

been reported (Wei et al., 2011). The hollow fiber and flat sheet

FO membranes not only achieved high water fluxes but also

superior rejection properties.

The overall energy utilized by the FO process has been re-

ported to be approximately 25e45% of the thermal energy

needed for multieffect distillation (Cath et al., 2009). FO has

the added capability for using heat at a much lower or higher

temperature than multieffect distillation processes. The FO

process can use heat as low as 40 �C and as high as 200e250 �C.A specific energy consumption of less than 0.25 kWh/m3 for

themembrane portion of FOwas reported by Cath et al. (2009).

However, after consideration of the energy consumption

associated with draw solution recovery, the specific energy

consumption was similar to an RO process (~3.5 kWh/m3).

In an innovative approach to reducing energy consump-

tion, Cath et al. (2009) used FO in combination with RO to form

a hybrid process (Cath et al., 2009). In this novel approach,

recycled water (tertiary treated effluent) was passed through

an FO system, with seawater used as the draw solution. The

seawater was diluted by the recycled water within the FO

process. The diluted seawater was then passed through a RO

system where the feed pressure requirement was lowered by

dilution of the seawater; hence, lower energy consumption

was obtained for the seawater desalination process. The

concentrate (brine) from the RO process was further treated

through a second stage FO process and the final seawater

brine was discharged to the ocean. By using a combination of

FO and RO, seawater desalination was performedwith a lower

energy consumption and the recycled water was simulta-

neously treated through two physical barriers (FO and RO)

(Cath et al., 2009). Using the same approach, Yangali-

Quintanilla et al. (2011) desalinated Red Sea water using FO

and RO. Secondary wastewater effluent was used as the feed

water and Red Sea water was used as the draw solution. The

system was reported to consume only 50% of the energy

(~1.5 kWh/m3) when compared to a high pressure seawater RO

system desalinating the same seawater. Although the FO/RO

process had lower energy consumption, a cost analysis sug-

gested that a minimum average flux of 10.5 L m�2 h�1 was

necessary to compete with a conventional RO process

(Yangali-Quintanilla et al., 2011).

In another study, Shaffer et al. (2012) reported a specific

energy consumption of 2.93 kWh/m3 at 50% recovery for an

integrated FO/RO process for producing water that can be

employed for agricultural purposes requiring low boron levels.

For the same conditions, a two-pass RO system will require

3.79 kWh/m3 of energy (Shaffer et al., 2012). However, the

combination of FO and RO will result in a substantially higher

membrane area requirement due to the operational flux for

the FOmembrane being low. For example, Shaffer et al. (2012)

reported a total membrane area of 972,000 m2 for the FO/RO

process at 50% total system recovery when compared to only

74,000 m2 for a two-pass RO process. Recently, Kim et al.

(2013a,b,c) utilized crystallization and RO for draw solution

recovery. The first step involved the FO process with the

diluted draw solution cooled in a crystallization unit. Due to

the crystallization step, the feed salinity to the RO recovery

system was reduced and a total energy consumption of

2.15 kWh/m3 for seawater desalination was obtained. The

total energy consumption was reported as the sum of cooling

energy in the crystallization process and pumping energy for

the RO process.

In a recent analysis, McGovern and Lienhard (2014)

compared the specific energy consumption of a two-pass RO

system for seawater desalination with forward osmosis. For

desalination of seawater with 35,000 mg/L of TDS at 50% re-

covery, the two-pass RO energy consumption was 3.0 kWh/m3

including ultrafiltration (pretreatment), 1st and the 2nd pass

RO. For the same conditions, the energy consumption for the

FO process was calculated to be 3.58 kWh/m3 with the draw

solution dilution and regeneration process consuming 0.10

and 3.48 kWh/m3, respectively (McGovern and Lienhard,

2014). Thus, for FO to be competitive with RO in terms of en-

ergy consumption, the regeneration process must be signifi-

cantly more efficient than RO. However, the FO process has

the advantage of lesser fouling propensity when compared to

RO due to the absence of a hydraulic driving pressure. The FO

process is also suitable for niche applications where the

salinity levels of the water that needs to be treated is higher

than the salinity that can handled with the RO process.

3. Thermal-based technologies

The principle of thermal-based desalination processes de-

pends on phase transition by energy addition or removal to

separate fresh water from saline water. The most important

thermal distillation processes are multistage flash (MSF),

multieffect distillation (MED) and vapor compression (VC) (Al-

Sahali and Ettouney, 2007). In recent years, modifications to

thermal-based desalination technologies have resulted in

increased process efficiency. Development has focused on

technologies that combine a thermal phase change with a

membrane. These technologies includemembrane distillation

and pervaporation. In order to reduce the energy consumption

of purely thermal processes, technologies such as humid-

ificationedehumidification and adsorption desalination have

been developed. These emerging technologies along with

their application in desalination is presented in this section. A

comparison of these thermal-based technologies is provided

in Table 2.

3.1. Membrane distillation

Membrane distillation (MD) is a thermally driven, membrane-

based process (El-Bourawi et al., 2006; Alkhudhiri et al., 2012;

Table 2 e Comparison of thermal-based technologies.

Technology Advantages Drawbacks Recovery range Feed waterquality

Treated waterquality

Energyconsumption

Cost impacts

Membrane

distillation

- Absence of applied

pressure.

- High rejection capacity

with the removal of non-

volatiles.

- Lower operating temper-

atures when used in

vacuum.

- Possibility of utilizing

plastic material to avoid

corrosion.

- Minimal effect of perfor-

mance due to feed salt

concentrations.

- Possibility of utilizing

different energy sources,

such as waste heat, solar

energy and geothermal.

- Only bench-scale

or pilot-scale

studies have been

performed.

- Lack of mem-

branes and mod-

ules designed

specifically for MD.

- Fouling of hydro-

phobic mem-

branes is an issue

when the mem-

brane is wetted.

Requires pretreat-

ment of feed water

source.

Up to 80%. No limitation on

feed water TDS.

Up to 300,000mg/

L TDS can be

treated.

Up to 100% rejection. 43 kWh/m3 without

waste heat (Walton

et al., 2004).

10.3 kWh/m3 with

waste heat.

Not known.

Technology still at

bench or

demonstration scale.

Costs depend on the

availability of a

waste heat source to

heat feed water

stream.

Humidificatione

dehumidification

- Simple operation and

maintenance.

- Can utilized solar energy

or waste heat source.

- High rejection capacity.

- Lower operating temper-

atures when compared to

conventional thermal

desalination.

- Ideal for small-scale

remote applications when

combined with solar

energy.

- Requires waste

heat or renewable

energy source for

cost-effective

desalination.

- High thermal en-

ergy consumption

(300e550 kWh/m3).

- Large footprint

requirement due

to humidifier and

dehumidifier

chambers.

- Optimization of

carrier gas flow

rate and feedwater

type is essential.

More than 97.5%. Up to seawater

salinities have

been evaluated.

Up to 100% rejection. 45.3 kWh/m3

(Mahmoud, 2011).

Less than $5/m3 for

the treatment of oil

and gas frac

flowback water.

Costs for other

applications are not

known.

Adsorption

desalination

- Low energy consumption.

- Utilization of low temper-

ature heat source or solar

heat.

- Stationary operation

without any moving

parts.

- Production of two effects,

distillation and cooling.

- Requires waste

heat or renewable

energy source for

cost-effective

desalination.

- Robustness of sil-

ica gel adsorber

beds is not known.

- Data available only

for pilot or

Up to 65% for seawater

desalination.

No limitation on

feed water TDS.

Up to 100% rejection. 1.38 kWh/m3 for

seawater

desalination (Ng

et al., 2013).

Not known.

Technology still at

demonstration scale.

water

research

75

(2015)164e187

174

demonstration-

scale

pro

jects.

Pervaporation

-Abse

nce

ofapplied

pressure.

-Highrejectionca

pacity.

-Loweroperating

temperatu

res.

-Low

flux.

-Perform

ance

de-

pendsonse

lection

ofmembrane

material.

-Lack

ofpilot-sc

ale

ordemonstration

scale

data.

Upto

100%.

Nolimitationon

feedwaterTDS.

Upto

100%

rejection.

Notknown.

Notknown.

Tech

nologystillat

bench

or

demonstrationsc

ale.

wat e r r e s e a r c h 7 5 ( 2 0 1 5 ) 1 6 4e1 8 7 175

Adham et al., 2013) that combines membrane technology and

evaporation processing in one unit. It involves the transport of

water vapor through the pores of hydrophobicmembranes via

the temperature difference across the membrane. This dif-

ference results in a vapor pressure differential, leading to the

transfer of the produced vapor through the hydrophobic

membrane to the condensation surface. In MD, performance

is not affected by the salinity of the feed water (Wang and

Chung, 2015). However, the permeate flux is strongly

affected by the feed temperature (Pangarkar et al., 2014).

Membranematerials that have been considered for the MD

process include polytetrafluoroethylene, polyvinylidene fluo-

ride, polyethylene, and polypropylene (Alkhudhiri et al., 2012).

Typical values for the porosity are 0.06e0.85; for the pore size,

0.2e1.0 mm; and for the thickness, 0.06e0.25 mm. For almost

three decades, MD has been considered an alternative to

conventional desalination technologies such as MSF and RO.

These two techniques involve high energy and high operating

pressure, respectively, which result in excessive operating

costs. MD offers the attraction of operation at atmospheric

pressure and low temperatures (30e90 �C), with the theoret-

ical ability to achieve 100% salt rejection.

Typical configurations utilized for MD are direct contact

membrane distillation (DCMD), air gap membrane distillation

(AGMD), vacuum membrane distillation (VMD) and sweeping

gas membrane distillation (SGMD) (Evans and Miller, 2002).

Different configurations of the MD process are utilized to

improve efficiency during desalination. In DCMD, the mem-

brane is in direct contact with the liquid phase. It is the

simplest configuration in MD and is considered to produce the

highest flux among the various configurations (Cath et al.,

2004). In AGMD, an air gap is present between the mem-

brane and condensate side. This configuration is considered to

exhibit the highest energy efficiency and is best suited for

applications with low energy availability (Meindersman et al.,

2006). In VMD, a vacuum pressure is maintained in the

permeate side and is best suited when volatiles need to be

removed in the feed water (Bandini et al., 1992). In the SGMD

configuration, a sweeping/stripping as is introduced in the

permeate side of the membrane (Basini et al., 1987). A review

of recent advances in membrane distillation with various

configurations and applications in brackish water desalina-

tion, seawater desalination, removal of small contaminant

molecules and recovery of valuable products has been pro-

vided in detail by Wang and Chung (2015). In this review,

different MD configurations along with hybrid configurations

have been discussed. Membranes based on polytetrafluoro-

ethylene (PTFE) dominate in MD modules due to its high hy-

drophobicity and resistance to harsh chemical conditions.

However, the cost of these membrane materials is prohibitive

for their large-scale implementation and difficulties associ-

ated with module design, specifically module sealing (Wang

and Chung, 2015). Thus, improvements in membrane mate-

rial are warranted for MD processes.

The selection of membrane material is critical for MD

processes. The selectedmembrane should be porous, must be

hydrophobic (should not be wetted by liquids) and must not

have capillary condensation inside the pores. Only vapor

should be transported through the pores and the membrane

must not alter vapor equilibrium of the process liquids

wat e r r e s e a r c h 7 5 ( 2 0 1 5 ) 1 6 4e1 8 7176

(Tomaszewska, 2000). One of the major drawbacks of the

technology is the low flux obtained (less than 8.5 Lm�2 h�1). In

an experimental study conducted for flux enhancement, Cath

et al. (2004) evaluated three hydrophobic membranes under

DCMD mode with a turbulent flow regime and a feed water

temperature of 40 �C. In this process, the feed water at

elevated temperature is in contact with one side of the

membrane and colder water is in direct contact with the

opposite side. The temperature difference and the solute

concentration induces the vapor pressure gradient for mass

transfer occurring in three steps: (1) diffusive transport from

the feed stream to the membrane interface, (2) combined

diffusive and convective transport for the vapors through the

membrane pores, and (3) condensation of the vapors on the

membrane interface on the product side of the membrane

(Cath et al., 2004). Using this configuration with synthetic

seawater resulted in a 99.9% salt rejection and a doubling of

the flux (up to 34 L m�2 h�1) when compared to conventional

DCMD processes. Due to the temperature gradient occurring

across themembrane, heat transfer across the boundary layer

often limits the rate of flux transfer in MD systems. In order to

improve mass transfer, Teoh et al. (2014) modified module

design using spacers and baffles for turbulence promotion.

Turbulence promoters aid in the increase of heat and mass

transfer coefficients resulting in increased flux transfer. Use of

baffles, external helix, inner helix and sieve designs have

shown a 11e49% increase in the flux at 75 �C when compared

to an unaltered module design (Teoh et al., 2014).

In an effort to enhance the overall feed water recovery of

RO processes, Qu et al. (2009) integrated DCMD with acceler-

ated precipitation softening to treat RO. The RO concentrate

was pH adjusted with sodium hydroxide along with calcite

seeding, followed by microfiltration to avoid crystal seeds

clogging the DCMDmodule. The overall feedwater recovery of

the desalination process was enhanced to 98.8% and the

DCMD flux declined only by 20% after 300 h of operation. In a

similar configuration, Mariah et al. (2006) integrated MD and

crystallization (MDC) to treat RO brine from seawater desali-

nation. In this approach, pure crystalline products were ob-

tained from RO concentrate as the vapor pressure driving

force during the process allowed a solution to be concentrated

up to saturation.

With SGMD, Evans and Miller (2002) evaluated Celgard

Liqui-Cel® Extra-Flow 2.5X8 membrane contactors with X-30

and X-40 hydrophobic hollow fiber membranes. Results

showed that the system was capable of producing low TDS

water, typically 10mg/L or less, from seawater using low grade

heat. However, large air flows were required to achieve sig-

nificant water yields, and the costs associated with trans-

porting this air were prohibitive. To overcome this barrier,

new and different contactor geometries are necessary to

achieve efficient contact with an extremely low pressure drop.

Second, the temperature limits of the membranes must be

increased. In the absence of these improvements, SGMD will

not be economically competitive.

Kim et al. (2013a,b,c) coupled MD with solar energy,

geothermal energy, or waste heat to reduce energy con-

sumption and cost. Even with the availability of a waste heat

source that can be utilized for the MD process, there are

limited studies on the MD process and a direct cost

comparison with conventional technologies (such as RO) is

not available. Moreover, the industry has not fully embraced

MD for several reasons: low water flux (i.e., productivity), en-

ergy efficiency and a shortage of long-term performance data

due to the wetting of the hydrophobic microporous mem-

brane (Mathioulakis et al., 2007). Wetting of the membrane

surface leads to accelerated deposition of organics on the

membrane surface resulting in robust pretreatment re-

quirements. Eventually, the pretreatment requirements in-

crease the cost of treatment. Innovative materials that offer

microporous membranes with desired porosity, hydropho-

bicity, low thermal conductivity, and low fouling are essential

to bring MD closer to commercialization. Opportunities,

therefore, beckon membrane researchers to improve the flux

in the process and increase its durability by fabricating highly

permeable superhydrophobic membranes and/or modifying

the MD module configurations.

There are several pilot and demonstration scale studies

being proposed to evaluate MD for brackish water treatment

and wastewater from oil and gas fields. In a study conducted

by through the National Center for Excellence in Desalination

(Australia), a MD system was employed for treatment of

brackish groundwater at Tjuntjunjarra, a remote corner state

of Western Australia where the primary contaminant is ni-

trate. At the Singapore Membrane Technology Center, MD is

being evaluated for treatment of water contaminated with oil

frompetroleum refineries consisting of awaste heat source. In

both these studies, a vacuum gap multi-effect membrane

distillation (V-MEMD) is being utilized. In the Australian study,

a 1 m3/d pilot systemwas utilized which can be operated with

a solar thermal or a waste heat source. Pretreatment consisted

of filtration depending on the water quality. With a feed water

conductivity of 28,000 mS/cm, a distillation quality of 1.8 mS/cm

was produced (Pryor, 2012). In MEMD, warm water passes

through a series of MD effects working at progressively lower

pressures (). This increases the gain output ratio (GOR) to the

level that it could deliver large units with a GOR of 10, com-

parable to commercially available multi-effect distillation

systems. Due to the vacuum, water boils at a much lower

temperature (50e80 �C). The specific energy consumption for

MD systems has been reported to be about 43 kWh/m3 of

water produced (Walton et al., 2004). However, this energy

consumption will be substantially lower when waste heat is

available to heat the source water that requires treatment.

Research in MD needs to be focused on preparation of

membrane material with more structured morphology to

facilitate mass transfer and enhance flux. In this effort,

Gethard et al. (2011) studied desalination using carbon nano-

tube enhanced MD. In this study, carbon nanotubes were

immobilized in the pores of a hydrophobic membrane which

favorably altered the wateremembrane interactions to pro-

mote vapor permeability while preventing liquid penetration

into the membrane pores. With the incorporation of carbon

nanotubes, the flux and salt rejection was enhanced by 1.85

and 15 times, respectively with seawater (Gethard et al., 2011).

3.2. Humidificationedehumidification

Humidificationedehumidification (HDH) is a distillation pro-

cess and is based on the increased ability of air to carry water

wat e r r e s e a r c h 7 5 ( 2 0 1 5 ) 1 6 4e1 8 7 177

vapor at higher temperatures (Kabeel et al., 2013). A heated air

flow (carrier gas) is brought in contact with the feedwater that

needs to be treated. The air extracts a certain quantity of vapor

in the humidification zone. Distilled water is recovered at the

dehumidification zone by maintaining contact of humid air

with a cooling surface causing the condensation of a part of

the vapor mixed with the air (Kabeel et al., 2013). The system

consists of a humidifier, a dehumidifier and a heater to heat

either the carrier gas or the feed water stream (Narayan et al.,

2009). Due to the high energy consumption associated with

this type of desalination technology, various modifications

and improvements have been evaluated. These innovations

have included the use of multi stage air heated cycle, me-

chanical compression driven systems, thermodynamic

balancing, systems with a common heat transfer wall and

hybrid systems with RO (Narayan et al., 2011).

Chafik (2003) used a solar collector to heat the air and

saturate it before further heating and humidification. This

cycle was repeated in four or five stages to increase the exit

humidity from 4.5% to 9.3% for a four stage system. However,

the gain out ratio (GOR) increased only by 9% (Narayan et al.,

2010). Narayan et al. (2011) utilized a variable pressure HDH

system to increase the humidity ratio in the humidifier alone

without changing the pressure in the dehumidifier. This

configuration leads to an increased vapor content of themoist

air at sub-atmospheric pressures, leading to higher efficiency.

Holst (2007) balanced the temperature difference in HDH

system by circulating air by natural convention. This config-

uration resulted in lower thermal energy consumption

(120 kWh/m3) for the HDH system. HDH with a common heat

transfer wall enhances heat recovery between the humidifier

and dehumidifier with an increased GOR and operation at

atmospheric pressure. This method of operation is known as

dewvaporation (Hamieh and Beckman, 2006; Beckman, 2008).

Although high energy recovery is possible with dewvapora-

tion, a large surface area is required for condensation to occur.

This will result in higher footprint requirement for dewvapo-

ration systems.

Further developments have included the use of reverse

osmosis systems to desalinate the brine from the humidifier

(Narayan et al., 2009). In this configuration, the HDH system is

operated using a thermal vapor compression system. The

carrier gas from the humidifier is compressed in a thermal

vapor compression system using a steam supply and then

sent to a dehumidifier. The dehumidified carrier gas is

expanded to recovery energy in the form of work which is

then utilized to operate a reverse osmosis system. This

configuration showed lower thermal energy consumption.

However, the availability of medium pressure steam is

essential.

3.3. Adsorption desalination

Adsorption desalination is a thermal-based technology which

employs low temperature waste heat sources or solar heat to

power the sorption cycle using a highly porous silica gel (Ng

et al., 2013). In this method, water vaporization occurs in an

evaporator followed by vapor adsorption/desorption onto sil-

ica gel and condensation at the condenser (Ghaffour et al.,

2014). The adsorption desalination cycles are operated in

batches in one or more pairs of reactors. In one reactor (one

half-cycle), the silica gel adsorbent is used to adsorb the vapor

generated in the evaporator. The saturated silica gel in

another bed (next half-cycle) is regenerated by low tempera-

ture heat source (typically 50e85 �C) or solar heat. The des-

orbed vapor is then condensed on tube surfaces of a

condenser. The major components involved are the evapo-

rator, the adsorber bedswith silica gel and the condenser. This

emerging desalination method produces high quality potable

water and cooling power with only one heat input source.

The silica gel adsorbent has a high affinity towards water

vapor due to the double bond surface forces that exist be-

tween the meso-porous silica gel and the water vapor. This

results in a high uptake of water vapor and regeneration by a

low temperature waste heat source whichwould otherwise be

purged into the atmosphere without utilization (Ng et al.,

2013). A solar powered system has been installed at the

demonstration-scale in Saudi Arabia and Singapore (Ng, 2014).

Another prototype has been installed in Singapore that uti-

lizes a waste heat source. Three large-scale systems are also

being planned for implementation in the Saudi Arabia (Amy

et al., 2013). Specific energy consumption <1.5 kWh/m3 has

been reported for seawater desalination using this technol-

ogy, which is substantially lower (with the presence of a waste

heat source) than seawater desalination using conventional

thermal-based and membrane-based technologies (Ng et al.,

2013).

3.4. Pervaporation

Pervaporation processes separate mixtures in contact with a

membrane via preferential removal of one component from

the mixture due to its higher affinity with, and/or faster

diffusion through the membrane. In desalination applica-

tions, pervaporation has the advantage to achieve near 100%

salt rejection with potential low energy consumption. It is a

combination of membrane permeation and evaporation. Per-

vaporation of an aqueous salt solution can be regarded as a

separation of pseudo-liquid mixture containing free water

molecules and bulkier hydrated ions formed in solution upon

dissociation of the salt in water (Kuznetsov et al., 2007).

Several membrane materials have been evaluated for the

process. Polyvinyl alcohol (PVA) membranes have been stud-

ied extensively as pervaporation material in various applica-

tions due to its excellent film-forming and highly hydrophilic

properties and high degrees of swelling due to the presence of

hydroxyl groups (Kittur et al., 2003). In other studies, hybrid

organic-inorganic membranes based on PVA, maleic acid and

silica have been utilized. Hybrid membranes showed higher

water flux and up to 99.9% salt rejection (Xie et al., 2010). The

introduction of silica nanoparticles in the polymer matrix

enhanced both the water flux and salt rejection due to

increased diffusion coefficients of water through the

membrane.

Polyetheramide-based polymer film of 40 mm thickness has

also been evaluated for pervaporation (Zwijnenberg et al.,

2005). The membrane in tubular configuration was used in a

direct solar membrane pervaporation process with seawater

as the feed. The retention of ions (sodium, chloride, calcium,

arsenic, boron and fluoride) was higher than RO (Zwijnenberg

wat e r r e s e a r c h 7 5 ( 2 0 1 5 ) 1 6 4e1 8 7178

et al., 2005). Results indicated that the measured fluxes were

independent of fouling and concentration up to 100 g/L of

solids. In another study, grafted poly (ethylene terephthalate)

films with styrene exhibited better selectivity to toluene when

compared to ungraftedmembranes for pervaporation (Khayet

et al., 2006). The process has also been evaluated using PEBAX

membranes with 100 mm thickness with an operating tem-

perature of 50 �C that allowed a water flux of 26.4 L m�2 h�1

(Hamouda et al., 2011). Pervaporation has also been imple-

mented for irrigation where a network of subsurface pipes is

filled with water that needs to be desalinated. The pipes are

lined with a unique DuPont material that allows water vapor

to diffuse through the pipe walls, while the contaminants are

retained within the pipes (DTI-r, 2012).

The main disadvantage of the process is the low water flux

achieved. Xie et al. (2010) showed that at low temperatures,

the salt concentration in the feed solution showed negligible

effects on water flux and diffusion coefficients. At high feed

temperatures (50e60 �C), flux and diffusivity of water

decreased with increasing salt concentration due to the

decreased water vapor pressure and water concentration at

the membrane surface. The feed water temperature is a

crucial parameter due to the increase in diffusivity and

reduction in viscosity that occurs on heating. In addition,

presence of vacuum and membrane thickness and inherent

permeability of the membrane polymer are important pa-

rameters determining the performance of the process.

4. Alternative technologies

Technologies discussed in this section include those that

operate using a mechanism different from the membrane or

thermal based technologies discussed in the previous sec-

tions. These alternative technologies are in the initial stages of

development and only a few pilot- or demonstration-scale

evaluations have been conducted. A comparison of these

desalination technologies is provided in Table 3.

4.1. Microbial desalination cell

Microbial desalination cells (MDCs) are based on the transfer

of ionic species from water in proportion to the current

generated by bacteria (Cao et al., 2009; He, 2012; Luo et al.,

2012). The two primary purposes of the MDCs are to

generate electricity and remove contaminants. MDCs consist

of three chambers, with an anion exchange membrane (AEM)

next to the anode, a cation exchange membrane (CEM) adja-

cent to the cathode, and a middle chamber between the

membranes filled with the water to be desalinated. When

current is generated by the bacteria growing on the anode,

positively charged ionic species are prevented from leaving

the anode by the AEM. Simultaneously, negatively charged

ionic species (such as Cl�) move from the middle chamber to

the anode. Simultaneously, as the cathode chamber protons

are consumed, positively charged species ionic species (such

as sodium ion) migrate from the middle chamber to the

cathode chamber (Cao et al., 2009). Thus, the loss of ionic

species from themiddle chamber results in the desalination of

the water without any feed pressure or draw solution and

electricity requirements. Instead, electricity is generated

while thewater is desalinated, a process similar to electrolysis

without the use of an external energy source (Cao et al., 2009).

The sludge generated during the process can be dewatered

and used as a substrate for the anode to enhance the stability

of desalination and electricity generation (Meng et al., 2014).

Initial proof-of-concept studies involved different initial

salt concentrations (5, 20 and 35 g/L) with acetate used as the

substrate for the bacteria (Cao et al., 2009). The MDC produced

a maximum of 31 W/m3, while at the same time removing

approximately 90% of the salt in a single desalination cycle. In

another study, Jacobson et al. (2011) desalinated artificial

seawater using a MDC; an increased removal of TDS (about

70%) was observed with an increase in hydraulic retention

time. It was determined that electricity generation was a

predominant factor in the removal of TDS and that other

factors such as water osmosis contributed to desalination.

Zhang et al. (2011) integrated MDC with forward osmosis for

wastewater treatment. In this approach, an osmotic microbial

fuel cell (OsMFC) was developed by using a FOmembrane as a

separator (absence of the middle chamber in a conventional

MDC). A NaCl solution or artificial seawater was used as the

draw solution and it was found that the OsMFC produced

more electricity than a conventional MDC in both batch and

continuous operation. Higher efficiency was attributed to

better proton transport with water flux through the FO

membrane. The OsMFC has applications in water reuse with

re-concentration of draw solutions using RO concentrate or in

seawater desalination using a MDC for further wastewater

treatment and desalination (Zhang et al., 2011).

Efforts to improve the efficiency of MDCs have resulted in

increased performance. In order to scale up MDCs, more

compact reactor designs are necessary (Yang et al., 2012).

Spacers can be used to minimize the reactor size without

adversely affecting performance. A single 1.5 mm expanded

plastic spacer produced a maximum power density (approxi-

mately 973 mW/m2) and was found to be similar to that of a

MDC with the cathode exposed to air without a spacer (Yang

et al., 2012). Use of a thinner spacer (1.3 mm) resulted in

reduced power by 33%. In order to recirculate the solutions

between the anode and cathode, a recirculation MDC (rMDC)

was designed and evaluated (Qu et al., 2012). This adaptation

avoided pH imbalances that could inhibit bacterial meta-

bolism. A salt solution (20 g/L NaCl) was reduced in salinity by

37% with the rMDC when compared to a standard MDC with

25% reduction in salinity (Qu et al., 2012). Stacked MDCs

resulted in efficient desalination of seawater when compared

to single chamber MDCs (Kim and Logan, 2011). When 4

stackedMDCswere used in series, the salinity of seawaterwas

reduced by 44%. The use of two additional stages resulted in

an increase in salt removal capacity by 98% (Kim and Logan,

2011). In another interesting application, a three chamber

MDC was used as a pretreatment to RO to reduce the salinity

of the feed water (Mehanna et al., 2010). A single cycle of

operation resulted in a 43% reduction in salinity and produc-

tion of 480 mW/m2. Thus, the use of MDC before RO results in

a reduction of salt concentration in the feed water with

reduced energy demand for the downstream RO process; at

the same time electrical power is produced (Mehanna et al.,

2010). Although MDCs show promise for desalination, the

Table 3 e Comparison of alternative technologies.

Technology Advantages Drawbacks Recovery range Feed water quality Treated waterquality

En y consumption Cost impacts

Microbial

desalination

cell

- Absence of applied

pressure.

- Absence of

external electricity

source.

- Low efficiency.

- Requires a carbon source.

- Only bench-scale studies

have been performed.

Recovery depends on

the hydraulic

retention time.

Up to seawater

salinity.

Up to 98% removal of

salinity with multi-

staged systems.

No e gy consumption.

How r, electricity

prod ion efficiency is

depe nt on several

facto

Not known. Technology still

at bench-scale level. Costs

will depend on scalability

and performance of

technology with real feed

water sources.

Capacitive

deionization

technologies

- Absence of applied

pressure.

- High rejection of

salt.

- More efficient for

low salinity feed

water sources

(TDS < 15,000 mg/

L).

- Polarity reversal

results in self-

cleaning of

electrodes.

- Efficiency of electrodes for

salt separation requires

optimization.

- Limited data available for

seawater desalination.

Up to 90%. TDS <15,000 mg/L. Up to 99% removal of

salt.

1.8 k /m3 for seawater

desa tion using a

com ation of

elect ialysis and

cont ous

elect eionization

(Siem s, 2014).

Not known. Costs will

depend on the feed water

TDS and associated energy

consumption.

Ion concentration

polarization

- Absence of applied

pressure.

- Absence of mem-

branes and fouling.

- High rejection of

salt and

microorganisms.

- Efficient for small-

scale desalination

modules.

- Can be operated

using battery

power in remote

locations.

- Limited data available.

- Applicable only for small-

scale systems.

- Scaling of microchannels

due to hardness ions

could be an issue.

Up to 50% for

seawater

desalination.

TDS <40,000 mg/L. More than 99%

removal of salt.

3.5 k /m3 for seawater

desa tion (Kim et al.,

2010

Not known. Technology still

at bench-scale level.

Clathrate

hydrates

- Low pressure

requirements.

- High rejection of

salt.

- Process has not been

evaluated on a continuous

basis.

- Separation of hydrates

from brine requires

optimization.

- Elimination of salt mole-

cules from hydrate cages.

Up to 100%. Up to seawater

salinity has been

evaluated.

Up to 100% removal

of salt.

Not wn. Estimated to be

sign ntly lower than RO

proc due to absence of

feed ssure requirement.

$0.46e$0.52/m3 for

seawater desalination.

water

research

75

(2015)164e187

179

erg

ner

eve

uct

nde

rs.

Wh

lina

bin

rod

inu

rod

en

Wh

lina

).

kno

ifica

ess

pre

wat e r r e s e a r c h 7 5 ( 2 0 1 5 ) 1 6 4e1 8 7180

process is still under development and only bench-scale

studies have been performed to date. Further pilot-scale sys-

tems need to be setup with a continuous mode of operation

with actual seawater or wastewater in order to prove the ef-

ficacy of this technology.

4.2. Capacitive deionization technologies

Although capacitive deionization (CDI) technology is not a

recent concept, several challenges exist for the identification

of an optimum material for electrode manufacture (Farmar

et al., 1997). CDI technology was developed as a nonpol-

luting, energy-efficient, and cost-effective alternative to

desalination technologies such as reverse osmosis and elec-

trodialysis (Welgemoed, 2005). In this technology, a saline

solution flows through an unrestricted capacitor-type module

consisting of numerous pairs of high-surface-area electrodes.

The electrode material, typically a carbon aerogel, has a high

specific surface area (400e1100 m2/g) and a very low electrical

resistivity (less than 40 mU cm). Anions and cations in solu-

tion are electrosorbed by the electric field upon polarization of

each electrode pair by a DC power source. After the adsorption

of ions, the saturated electrode undergoes regeneration by

desorption of the adsorbed ions under zero electrical potential

or reverse electric field (Seo et al., 2010). Thus, the adsorption

ability of the electrode is the key parameter for the perfor-

mance of CDI technology.

4.2.1. Membrane-based systemsWhen a potential is applied to CDI electrodes, counter-ions are

attracted onto the electrode surface; simultaneously co-ions

are expelled from the counter electrode (Kim and Choi,

2010). This leads to higher energy consumption and a lower

operation efficiency because of themobility of unwanted ions.

Recently, a modification of capacitive deionization has resul-

ted in higher recovery and efficiency in a membrane-CDI

(MCDI) technology (Kim and Choi, 2010; Biesheuvel and van

der Wal, 2010) where ion-exchange membranes are used for

selective transport of ions to the electrodes. This advance has

resulted in higher efficiency and lower energy consumption.

4.2.2. Flow through systemsIn another method, called flow-through electrode CD system,

the feedwater flows through the electrodes, instead of flowing

between them (RWL, 2012). This method is four to ten times

faster than conventional CDI, is more energy efficient, and

requires no membrane components. The system also uses a

new electrode material called hierarchical carbon aerogel

monoliths in lieu of carbon aerogels. The porous carbon ma-

terial eliminates the limitations typically associated with

conventional capacitive deionization. It also has less separa-

tion between electrodes to further increase the system's per-

formance. The entire process can be executed in only the time

needed to charge the electrodes (RWL, 2012).

4.2.3. Hybrid systemsEnergy consumption as low as 0.1 kWh/m3 has been reported

in using CDI for brackish water treatment (Welgemoed, 2005).

For seawater desalination, an energy consumption of

1.8 kWh/m3 using a combination of ED and continuous

electrodeionization (CEDI) was recently reported (Siemens,

2014). In the hybrid approach, an electric field is used to

draw sodium and chloride ions across ion-exchange mem-

branes. As the water itself does not pass through the mem-

branes, the process can be operated at lower pressure and

lower energy consumption. Seawater is pretreated with a self-

cleaning disk filter, followed by UF modules. The EDeCEDI

system consists of ED units arranged in series to remove high

concentrations of salt, followed by CEDI units arranged in

parallel to remove smaller amounts of salt. Besides energy

savings, other claimed advantages of the EDeCEDI technology

include lower vibration and noise levels, improved safety, and

minimal pre- and post-treatment (Siemens, 2014).

In another application, the Voltea process, the technology

combines ED and CDI (Voltea, 2012). A three-step process is

utilized, with the water flowing in a cell containing positively

and negatively charged electrodes. The electrode surfaces are

covered with ion-selective membranes, so ions in the feed

water are attracted to the oppositely charged electrodes, pass

through the membrane, and finally accumulate within the

porous electrode structure. Up to 99% salt rejection has been

reported using the process. When the electrodes become

saturated, their polarity is reversed. The process is estimated

to use less than 1.0 kWh/m3 when removing 3000 mg/L of salt

from water. The system can operate at 90% recovery and can

be equipped with an energy recovery system to reuse the

energy stored in ions on the electrodes.

4.2.4. Entropy battery systemsSimilar to the CDI technique, a desalination battery has been

evaluated (Pasta et al., 2012). The system was previously re-

ported as a “mixing entropy battery” (LaMantia et al., 2011; Lee

et al., 2014). Instead of storing charge in the electrical double

layer at the surface of the electrode, it is held in the chemical

bonds, which is the bulk of the electrode. Battery electrodes

offer higher specific capacity and lower self-discharge than

capacitive electrodes (Pasta et al., 2012). The desalination bat-

tery operates by performing cycles in reverse when compared

to the mixing entropy battery. Rather than generating elec-

tricity from salinity differences, as in the mixing entropy bat-

tery, the desalination battery uses an electrical energy input to

extract sodium and chloride ions from seawater to generate

freshwater. The systemconsists of a cationic sodium insertion

electrode and a chloride capturing anionic electrode. It utilizes

a Na2exMn5O10 nanorod positive electrode with Ag/AgCl

negative electrode (Pasta et al., 2012). A four step charge/

dischargeprocessallows theseelectrodes to separate seawater

into fresh water and brine streams. In the first step, the fully

charged electrodes are immersed in seawater. A constant

current is then applied in order to remove the ions from the

solution. In the second step, the freshwater solution in the cell

is extracted and then replaced with additional seawater. The

electrodes are then recharged in this solution, releasing ions

and creating the brine in the third step. In the fourth step, the

brine solution is replaced with fresh seawater and the cycles

are repeated (La Mantia et al., 2011). Steps 1e2 results in the

production of fresh water while steps 3e4 in brine production.

Wire-based systemsRecently, desalinationwith capacitivewire-

based technology with anode/cathode wire pairs was con-

structed from coating a thin porous carbon electrode layer on

wat e r r e s e a r c h 7 5 ( 2 0 1 5 ) 1 6 4e1 8 7 181

top of electrically conducting wires (Porada et al., 2012). Using

this novel approach, feed water with a salinity of 20 mM was

reduced by a factor of 3e4. By alternately dipping an array of

electrode pairs in freshwater with and in brine without an

applied cell voltage, an ion adsorption/desorption cycle is

created. The wires used are rigid graphite rods having a length

of about 40 times their thickness, chosen specifically for their

property of being inert, cheap and highly conductive. When a

voltage difference is applied between the wires, the cathode

adsorbs cations and the anode adsorbs anions. During the

process, ions are temporarily stored in the electrical double

layers (EDLs) formed within the micropores inside the carbon

particles that constitute the electrode. The voltage required

can be produced by solar panels and recoveries up to 50% can

be achieved by theprocess. Theprocess alsohas the advantage

of producing fresh water continuously as opposed to being

produced intermittently by CDI, thus avoiding mixing of

freshwater and untreated feed water. The wires can be placed

close toeachotherwithout requiringa spacer layer inbetween,

thus reducing pressure drops and fouling (Porada et al., 2012).

4.3. Ion concentration polarization

Ion concentration polarization has been utilized to desalinate

seawater using an energy-efficient process (Kim et al., 2010). In

this process, micro- and nanofluidics in combination with ion

concentration polarization are used to desalinate seawater.

Ion concentration polarization is a fundamental transport

mechanism that occurs when an ionic current is passed

through an ion-selective membrane. But, in the newly devel-

oped process, no membranes are utilized. Poly-

dimethylsiloxane (PDMS) microfluidic chips with perm-

selective nanojunctions are used for desalination (Kim et al.,

2010). An electrical potential is used to create a repulsion

zone that acts as a membrane separating charged ions, bac-

teria, viruses, and microbes from seawater flowing through a

500 � 100 mm microchannel. Water flows through the micro-

channel tangential to a nanochannel where the voltage is

applied. The resulting force creates a repulsion zone and the

stream splits into two smaller channels at the nanojunction.

The two streams created are the treated water and concen-

trate.More than 99% salt rejection and 50% recovery have been

reported using this process (Kim et al., 2010). When seawater

was used for the experiments, a depletion zone was created

within 1 s to divert charged ions into the brine stream. The ICP

layer acted as a virtual barrier for any charged particles, both

positiveandnegative, includingbiomoleculesdue to theirnon-

zero zetapotential (Kimet al., 2010). Thus, theprocess removes

both small salt ionsand largemicroorganisms simultaneously.

The process worked efficiently for 1 h without any clogging of

the microchannels. The ion concentration polarization pro-

cess has been reported to consume approximately

3.5e3.75 kWh/m3 of energy (Kim et al., 2010). The process will

ultimately be most appropriate for small-scale systems, with

the possibility of battery-powered operation.

4.4. Clathrate hydrates

Clathrate hydrates are crystalline inclusion compounds of

water (majority species-water molecules) and a guest

molecule (minority species-gas molecules) that form sponta-

neously at conditions of temperature and pressure particular

to each guestmolecule (Bradshaw et al., 2008; Park et al., 2011).

Typically, the temperatures at which clathrate hydrates are

stable are slightly above the freezing point of water, although

certain guest molecules stabilize hydrates at nearly ambient

temperature (Bradshaw et al., 2008). Three types of crystalline

structures that clathrate hydrates possess are structure I,

structure II and structure H. Each of the structures contains

cage-like sub structures that are formed by water molecules

and enclathrate the guest molecule (Bradshaw et al., 2008).

Hydrogen bonding is the primary mechanism of interaction

between the waterewater molecules while van der Waals

forces are responsible for the stabilization of the guest mole-

cules. After occupation of a sufficient number of cages, a

thermodynamically stable crystalline unit cell structure is

formed. These are referred to as gas hydrates and often

contain guest molecules that exist in the gas phase at stan-

dard temperature and pressure. Common types of naturally

occurring gas hydrates are methane, carbon dioxide,

hydrogen sulfide, ethane, ethylene and propane (Bradshaw

et al., 2008).

The use of clathrate hydrates to facilitate freeze desalina-

tion of saline water has been investigated since the 1960's.Processes based on propane and a variety of hydrates were

investigated for kinetics and separations in continuous flow

systems (Knox et al., 1961; Barduhn et al., 1962). In the 1990's,the Bureau of Reclamation commissioned work from Thermal

Energy Systems, Inc., to explore the feasibility of this tech-

nology and construct of pilot plant facilities in Hawaii and San

Diego (McCormack and Niblock, 2000, 1998; McCormack and

Andersen, 1995). In the Thermal Energy Systems study, the

hydrochlorofluorocarbon (HCFC) refrigerant, R141b, was

pressurized in a tank of seawater at hydrate forming condi-

tions (100 psig and 5 �C). A complex slurry of hydrates and

brine was formed. In order to obtain potable water, the hy-

drates need to be separated from the brine. The hydrates

formed were small in size and dendritic in nature with a high

amount of salt entrapped in the interstitial spaces. In order to

remove the salt, filtration was utilized to remove the smallest

hydrates and a wash column to remove excess salt. The wash

step results in the production of potable water quality and an

estimated cost of $0.46e$0.52/m3 was reported (McCormack

and Niblock, 2000).

In another study (Bradshaw et al., 2008), the problem of

interstitial salt formation and the necessity of using wash

columns was eliminated. Novel hydrate formers, utilizing

R141b and HFC-32, resulted in hydrate formation significantly

above the freezing point of water. Additives were utilized to

inhibit dendritic hydrate growth and thus minimize intersti-

tial salt entrapment. The rate of R141b hydrate formation in

the presence of a heat exchange fluid depended on the degree

of supercooling and the use of perfluorocarbon heat exchange

fluid assisted separation of R141b hydrates from the brine. The

use of clathrate hydrates shows promise for desalination with

a production cost similar to commercially available desalina-

tion technologies but the process is yet to be operated on a

completely functional level and there are no continuously

operating facilities, even at a pilot scale. Drawbacks related to

the control of hydrate nucleation, hydrate size and

wat e r r e s e a r c h 7 5 ( 2 0 1 5 ) 1 6 4e1 8 7182

morphology, agglomeration, amount of entrapped salt, and

the efficient recovery of hydrates from the brine is necessary

for process optimization.

5. Application potential of emergingtechnologies in the treatment of challengingwastewater sources

A combination of two technologies (hybrid) has shown to be

more efficient than the utilization of a single technology by its

own. Several hybrid configurations are currently being eval-

uated for treatment of challenging wastewater sources. These

feed water sources include: flowback and produced water

from the upstream oil and gas industry; cooling tower blow

down water or refinery wastewater from the downstream oil

and gas industry and power industry; leachate and acid mine

drainage water treatment from the mining industry; high or-

ganics laden wastewater during the production of syngas and

biofuels; and metals laden wastewater in the metals

manufacturing industry. All of these industrial sectors require

potable water for various processes and applications. An

integration of emerging desalination technologies will not

only provide a means to treat these challenging wastewater

sources, but also potentially treat wastewater at higher feed

water recoveries with low operational and maintenance costs

associated with electricity consumption and cleaning chem-

icals usage.

A few hybrid configurations that can be utilized for the

treatment of various industrial wastewater sources are shown

in Fig. 1. In configuration (A), a forward osmosis system is

combinedwith a reverse osmosis system for treatment of high

fouling wastewaters (Holloway et al., 2007; Hickenbottom

et al., 2013). Since applied pressure is absent in forward

Fig. 1 e Application of hybrid technologies for tr

osmosis, deposition of foulants on the membrane is lower

leading to the elimination of pretreatment. Due to lack of

applied pressure, osmotic cleaning using a low salinity stream

in the draw solution side will result in transport of water from

the draw solution side to the feed solution side (Chung et al.,

2012). This will dislodge loosely held foulant deposits on the

membrane surface leading to more efficient cleaning. Con-

centration of the draw solution is accomplished using a con-

ventional reverse osmosis system. Due to the maximum feed

pressure limit associated with the reverse osmosis, the hybrid

configuration of FO and RO can only be utilized for treatment

of low salinity feed water streams. For feed water streams

with a TDS >40,000 mg/L, draw solution recovery using a

gaseous mixture of NH3/CO2 could be utilized. In this case,

additional energy requirements for the recovery of the draw

solution using heat or other thermal means must be consid-

ered. This configuration will be especially applicable for the

treatment of flowback water in oil and gas industries if

beneficial reuse is desired. The treatedwater can be reused for

beneficial reuse purposes, such as feed water for boilers or

irrigation.

In configuration (B), a forward osmosis system is combined

with a membrane distillation system. The membrane distil-

lation system is used for concentration of the draw solution

(Wang et al., 2011; Xie et al., 2013). Based on the salinity of the

feed water stream, various draw solutions can be used. Since

the membrane distillation system performance is not

restricted by salinity, this configuration can be used for

treatment of high salinity wastewater streams. A typical

application would include treatment of flowback or produced

water in oil and gas industry (USDOE, 2014). This hybrid

configuration promises minimal energy requirements when a

waste heat source is available to heat the draw solution and

re-concentrate using membrane distillation.

eatment of challenging wastewater sources.

wat e r r e s e a r c h 7 5 ( 2 0 1 5 ) 1 6 4e1 8 7 183

In configuration (C), membrane distillation is used for

treatment of concentrate of RO systems (Camacho et al., 2013;

Dow et al., 2008). In this configuration, the RO concentrate

may require pretreatment to remove organics and scale

forming ions present to reduce fouling and scaling of the

membrane in the membrane distillation system. Further

treatment of the RO concentrate using a membrane distilla-

tion system can reduce the brine volume generated and in-

crease the overall feed water recovery of the desalination

process. When zero liquid discharge is desired, the final brine

from the membrane distillation system can be sent to a brine

crystallizer. Utilization of a membrane distillation system

prior to the crystallizer can reduce costs substantially due to

lower brine volume treatment requirement in the crystallizer.

Several oil and gas facilities currently operate reverse osmosis

system and the feed water recovery is restricted due to the

salinity of the feed water stream (Adham, 2013). Using a

combination of RO and MD can potentially result in high feed

water recoveries while reducing energy costs required for

concentrate treatment.

In configuration (D), flowback water from the oil and gas

industry is pretreated and fed to a microbial fuel cell system

(Barole and Tsouris, 2013; Barole, 2014). In this system, elec-

tricity is generated which is then utilized for desalination in a

capacitive deionization system. The microbial fuel cell is also

utilized for removal of biodegradable organics in the feed

water and the second step utilizes a desalination step to

remove dissolved solids. This hybrid configuration is energy

self-sufficient by generating electricity on-site. The microor-

ganisms present in the fuel cell anode chamber degrade the

organics present in the feed water and hence this step acts as

a pretreatment to the capacitive deionization system. The

advantage of using this hybrid configuration is the absence of

an external electricity source since the electricity required by

the capacitive deionization system is provided by the micro-

bial fuel cell. Additional energy generated through the mi-

crobial fuel cell can be utilized for other applications on-site.

The primary drawback with this configuration is the TDS limit

of the feed water. Capacitive deionization systems are effi-

cient only for low TDS (<15,000 mg/L). Also, the scalability of

this hybrid configuration is not known and only bench-scale

experimental data are currently available.

6. Conclusions

Several new technologies have been developed for desalina-

tion in recent years. Certain emerging membrane-based

technologies, such as nanocomposite membranes and

closed circuit desalination, show substantial promise for en-

ergy reduction and have been recently commercialized.

Although technologies based on aquaporins and nanotubes

show promise for high permeability and minimum energy

consumption, these technologies are not developed to the

point of commercialization and further studies are required at

a larger scale to determine their sustainable operation. In

addition, technologies based on nanotubes will still be

restricted due to the thermodynamic and osmotic pressure

restrictions and it is not clear if significant energy reduction is

possible beyond the minimum theoretical energy

consumption. Membrane-based technologies, such as for-

ward osmosis, can achieve lower energy consumption only if

a waste heat source is available for regeneration of the draw

solution. Other process configurations with FO and RO for

simultaneously treating wastewater and seawater shows

promise, but water quality parameters related to mixing of

emerging contaminants with the treated seawater must be

considered. Technologies related to microbial cells do not

require an external energy source but the efficiency of these

technologies for full-scale desalination applications is ques-

tionable. Among the thermal-based technologies, membrane

distillation is innovative and shows the most commercial

opportunity if the operational flux can be improved. Among

alternative technologies evaluated in this review, capacitive

deionization based technologies can potentially provide lower

energy consumption as well as superior performance when

compared to other desalination technologies, but require

operation in combination with other technologies for sus-

tainable performance.

Acknowledgments

The authors would like to thank the WateReuse Research

Foundation for project funding (Project # WateReuse-11-04),

the Foundation's project advisory committee and its project

officer, Kristan Cwalina.

r e f e r e n c e s

Achilli, A., Cath, T.Y., Childress, A.E., 2010. Selection of inorganic-based draw solutions for forward osmosis applications. J.Memb. Sci. 364, 233e241.

Adham, S., 2013. Membrane distillation using low grade wasteheat. In: Proceedings of the American Membrane TechnologyAssociation (AMTA) Membrane Technology Conference, SanAntonio, Texas.

Adham, S., Hussain, A., Matar, J.M., Dores, R., Janson, A., 2013.Application of membrane distillation for desalting brines fromthermal desalination plants. Desalination 314, 101e108.

Agre, P., 2003. Aquaporin Water Channels, Nobel Lecture.Stockholm, Sweden.

Ahadian, S., Kawazoe, Y., 2009. An artificial intelligence approachfor modeling and prediction of water diffusion inside a carbonnanotube. Nanoscale Res. Lett. 1054e1058.

Ahn, C.H., Baek, Y., Lee, C., Kim, S.O., Kim, S., Lee, S., Kim, S.-H.,Bae, S.S., Park, J., Yoon, J., 2012. Carbon nanotube-basedmembranes: fabrication and application to desalination. J. Ind.Eng. Chem. 18, 1551e1559.

Al-Sahali, M., Ettouney, H., 2007. Developments in thermaldesalination processes: design, energy, and costing aspects.Desalination 214, 227e240.

Alkhudhiri, A., Darwish, N., Hilal, N., 2012. Membrane distillation:a comprehensive review. Desalination 287, 2e18.

Amy, G., Ghaffour, N., Nunes, S., Ng, K.C., 2013. Recentdevelopments in low-energy desalination technologies:forward osmosis, membrane distillation, and adsorptiondesalination. In: Presentation at the WateReuse ResearchConference, Phoenix, Arizona. Available at: http://www.watereuse.org/sites/default/files/u3/WateReuse%20(PHX)%202013-Gary%20Amy.pdf (last accessed 21.11.14.).

wat e r r e s e a r c h 7 5 ( 2 0 1 5 ) 1 6 4e1 8 7184

Bandini, S., Gostoli, C., Sarti, G.C., 1992. Separation efficiency invacuum membrane distillation. J. Memb. Sci. 73, 217e229.

Barduhn, A.J., Towlson, H.E., Lee, Y.C., 1962. The properties ofsome new gas hydrates. J. AICHE 176, 8.

Barole, A., 2014. Sustainability in water treatment and reuse:bioelectrochemical systems for synergistic salt and BODremoval. In: Proceedings of the 29th Annual WateReuseSymposium, Dallas, Texas.

Barole, A., Tsouris, C., 2013. US Patent 8, 597, 513 B2, Microbial fuelcell treatment of fuel processing wastewater.

Bartman, A., Zhu, A., Christofides, P.D., Cohen, Y., 2010.Minimizing energy consumption in reverse osmosismembrane desalination using optimization-based control. J.Process Control 20, 1261e1269.

Basini, L., D'Angelo, G., Gobbi, M., Sarti, G.C., Gostoli, C., 1987. Adesalination process through sweeping gas membranedistillation. Desalination 64, 245e257.

Beckman, J.R., 2008. Dewvaporation Desalination 5,000 Gallon PerDay Pilot Plant. Desalination and Water Purification Research(DWPR) Final Report No. 120, Available online: http://www.usbr.gov/research/AWT/reportpdfs/report120.pdf (lastaccessed 19.11.14.).

Biesheuvel, P.M., van der Wal, A., 2010. Membrane capacitivedeionization. J. Memb. Sci. 346, 256e262.

Bowen, W.R., 2006. Biomimetic separations e learning from theearly development of biological membranes. Desalination 199,225e227.

Bradshaw, R.W., Greathouse, J.A., Cygan, R.T., Simmons, B.A.,Dedrick, D.E., Majzoub, E.H., 2008. Desalination UtilizingClathrate Hydrates (LDRD Final Report). Sandia NationalLaboratories. SAND2007-6565, January 2008.

Camacho, L.M., Dum�ee, L., Zhang, J., Li, J., Duke, M., Gomez, J.,Gray, S., 2013. Advances in membrane distillation for waterdesalination and purification applications. Water 5, 94e196.

Cao, X., Huang, Liang, P., Xiao, K., Zhou, Y., Zhang, X., Logan, B.E.,2009. A new method for water desalination using microbialdesalination cells. Environ. Sci. Technol. 43, 7148e7152.

Cath, T.Y., Adams, V.D., Childress, A.E., 2004. Experimental studyof desalination using direct contact membrane distillation: anew approach to flux enhancement. J. Memb. Sci. 228, 5e16.

Cath, T., Childress, A., Elimelech, M., 2006. Forward osmosis:principles, applications, and recent developments. J. Memb.Sci. 281, 70e87.

Cath, Tzahi Y., Drewes, J.E., Lundin, C.D., 2009. A Novel HybridForward Osmosis Process for Drinking Water AugmentationUsing Impaired Water and Saline Water Resources. WaterResearch Foundation and Arsenic Water TechnologyPartnership.

Chafik, E., 2003. A new type of seawater desalination plants usingsolar energy. Desalination 156, 333e348.

Choi, W., Choi, J., Bang, J., Lee, J.H., 2013. Layer-by-layer assemblyof graphene oxide nanosheets on polyamide membrane fordurable reverse-osmosis applications. ACS Appl. Mater.Interfaces 5, 12510e12519.

Chung, T., Zhang, S., Wang, K.Y., Su, J., Ling, M.M., 2012. Forwardosmosis processes: yesterday, today and tomorrow.Desalination 287, 78e81.

Coday, B.D., Heil, D.M., Xu, P., Cath, T.Y., 2013. Effects oftransmembrane hydraulic pressure on performance offorward osmosis membranes. Environ. Sci. Technol. 47,2386e2393.

Corry, B., 2008. Designing carbon nanotube membranes forefficient water desalination. J. Phys. Chem. B 112, 1427e1434.

Dow, N., Zhang, J., Duke, M., Li, J., Gray, S.R., Ostarcevic, E., 2008.Membrane Distillation of Brine Wastes. Water QualityResearch Australia (WQRA) Final Report No. 63.

Drak, A., Adato, M., 2014. Energy recovery consideration inbrackish water desalination. Desalination 339, 34e39.

DTI-r, 2012. Subsurface Vapor Transfer Irrigation. Available at:http://launch.org/forum/1/water/innovators/5/subsurface-vapor-transfer-irrigation/profile (last accessed 06.09.12.).

Efraty, A., Barak, R.N., Gal, Z., 2011. Close circuit desalination e anew low energy high recovery technology without energyrecovery. Desalination Water Treat. 31, 95e101.

El-Bourawi, M.S., Ding, Z., Ma, R., Khayet, M., 2006. A frameworkfor better understanding membrane distillation separationprocess. J. Memb. Sci. 285, 4e29.

Elimelech, M., Phillip, W.A., 2011. The future of seawaterdesalination: energy, technology, and the Environment.Science 333, 712e717.

Evans, L.R., Miller, J.E., 2002. Sweeping Gas MembraneDesalination Using Commercial Hydrophobic Hollow FiberMembranes. Sandia National Labs Report, SAND 2002-0138.

Fang, W., Wang, R., Chou, S., Setiawan, L., Fane, A.G., 2012.Composite forward osmosis hollow fiber membranes:integration of RO- and NF-like selective layers to enhancemembrane properties of anti-scaling and anti-internalconcentration polarization. J. Memb. Sci. 394, 140e150.

Farmar, J.C., Tran, T.D., Richardson, J.H., Fix, D.V., May, S.C.,Thomson, S.L., 1997. The Application of Carbon AerogelElectrodes to Desalinate and Waste Treatment. LawrenceLivermore National Laboratory. Report No. 231717.

Gai, J.G., Gong, X.L., 2014. Zero internal concentration polarizationFO membrane: functionalized graphene. J. Mater. Chem. A 2,425e429.

Gai, J.G., Gong, X.L., Wang, W.W., Zhang, X., Kang, W.L., 2014. Anultrafast water transport forward osmosis membrane: porousgraphene. J. Mater. Chem. A 2, 4023e4028.

Gao, L., Rahardianto, A., Gu, H., Christofides, P.D., Cohen, Y., 2014.Energy-optimal control of RO desalination. Ind. Eng. Chem.Res. 53, 7409e7420.

Gethard, K., Sae-Khow, O., Mitra, S., 2011. Water desalinationusing carbon-nanotube-enhanced membrane distillation. ACSAppl. Mater. Interfaces 3, 110e114.

Ghaffour, N., Lattemann, S., Missimer, T., Ng, K.C., Sinha, S.,Amy, G., 2014. Renewable energy-driven innovative energy-efficient desalination technologies. Appl. Energy 136,1155e1165.

Greenlee, L.F., Lawler, D.F., Freeman, B.D., Marrot, B., Moulin, P.,2009. Reverse osmosis desalination: water sources,technology, and today's challenges. Water Res. 43, 2317e2348.

Hameeteman, E., 2013. Future Water (in)Security: Facts, Figures,and Predictions. Global Water Intelligence Report. Available at:http://www.gwiwater.org/sites/default/files/pub/FUTURE%20WATER%20(IN)SECURITY.pdf (last accessed 10.02.15.).

Hamieh, B.M., Beckman, T.R., 2006. Seawater desalination usingdewvaporation technique: experimental and enhancementwork with economic analysis. Desalination 195, 14e25.

Hamouda, S.B., Boubakri, A., Nguyen, Q.T., Amor, M.B., 2011.PEBAX membranes for water desalination by pervaporationprocess. High. Perform. Polym. 23, 170e173.

He, Z., 2012. One more function for microbial fuel cells in treatingwastewater: producing high-quality water. CHEMIK 66, 3e10.

Hickenbottom, K.L., Hancock, N.T., Hutchings, N.R.,Appleton, E.W., Beaudry, E.G., Xu, P., Cath, T.Y., 2013. Forwardosmosis treatment of drilling mud and fracturing wastewaterfrom oil and gas operations. Desalination 312, 60e66.

Hinds, B.J., Chopra, N., Rantell, T., Andrews, R., Gavalas, V.,Bachas, L.G., 2004. Aligned multiwalled carbon nanotubemembranes. Science 303, 62e65.

Hilder, T.A., Gordon, D., Chung, S., 2009. Salt rejection and watertransport through boron nitride nanotubes. Small 5,2183e2190.

Hofs, B., Schurer, R., Harmsen, D.J.H., Ceccarelli, C.,Beerendonk, E.F., Cornelissen, E.R., 2013. Characterization and

wat e r r e s e a r c h 7 5 ( 2 0 1 5 ) 1 6 4e1 8 7 185

performance of a commercial thin film nanocompositeseawater reverse osmosis membrane and comparison with athin film composite. J. Memb. Sci. 446, 68e78.

Holloway, R.W., Childress, A.E., Dennett, K.E., Cath, T.Y., 2007.Forward osmosis for concentration of anaerobic digestercentrate. Water Res. 41, 4005e4014.

Holst, M.H., 2007. Solar thermal desalination using the multipleeffect humidification method. In: Solar Desalination for the21st Century. Springer, Dordrecht.

Holt, J.K., Park, H.G., Wang, Y.M., Stadermann, M.,Artyukhin, A.B., Grigoropoulos, C.P., Noy, A., Bakajin, O., 2006.Fast mass transport through sub-2-nanometer carbonnanotubes. Science 312, 1034e1037.

Hoover, L.A., Phillip, W.A., Tiraferri, A., Yip, N.Y., Elimelech, M.,2011. Forward with osmosis: emerging applications for greatersustainability. Environ. Sci. Technol. 45, 9824e9830.

Hummer, G., Rasaiah, J.C., Noworyta, J.P., 2001. Water conductionthrough the hydrophobic channel of a carbon nanotube.Nature 414, 188e190.

Humplik, T., Lee, J., O'Hern, S.C., Fellman, B.A., Baig, M.A.,Hassan, S.F., Atieh, M.A., Rahman, F., Laoui, T., Karnik, R.,Wang, E.N., 2011. Nanostructured materials for waterdesalination. Nanotechnology 22, 292001.

Jacobson, K.S., Drew, D.M., He, Z., 2011. Use of a liter-scalemicrobial desalination cell as a platform to studybioelectrochemical desalination with salt solution or artificialseawater. Environ. Sci. Technol. 45, 4652e4657.

Jeong, B.H., Hoek, E.M.V., Yan, Y., Subramani, A., Huang, X.,Hurwitz, G., Ghosh, A.K., Jawor, A., 2007. Interfacialpolymerization of thing film nanocomposites: a newconcept for reverse osmosis membranes. J. Memb. Sci. 294,1e7.

Jessop, P.G., Heldebrant, D.J., Li, X., Eckert, C.A., Liotta, C.L., 2005.Green chemistry: reversible nonpolar-to-polar solvent. Nature436, 1102.

Jessop, P.G., Mercer, S.M., Heldebrant, D.J., 2012. CO2- triggeredswitchable solvents, surfactants, and other materials. EnergyEnviron. Sci. 5, 7240e7253.

Joshi, R.K., Carbone, P., Wang, F.C., Kravets, Su, Y., Grigorieva, I.V.,Wu, H.A., Geim, A.K., Nair, R.R., 2014. Precise and ultrafastmolecular sieving through graphene oxide membranes.Science 343, 752e754.

Kabeel, A.E., Hamed, M.H., Omara, Z.M., Sharshir, S.W., 2013.Water desalination using a humidification-dehumidificationtechnique e a detailed review. Nat. Resour. 4, 286e305.

Kalra, A., Garde, S., Hummer, G., 2003. Osmotic water transportthrough carbon nanotube membranes. Proc. Natl. Acad. Sci.100, 10175e10180.

Kaufman, Y., Berman, A., Freger, V., 2010. Supported lipid bilayermembranes for water purification by reverse osmosis.Langmuir 26, 7388e7395.

Khayet, M., Nasef, M.M., Mengual, J.I., 2006. Application of poly(ethylene terephthalate)-graft-polystyrene membrane inpervaporation. Desalination 193, 109e118.

Kim, Y.J., Choi, J.H., 2010. Enhanced desalination efficiency incapacitive deionization with an ion-selective membrane. Sep.Purif. Technol. 71, 70e75.

Kim, Y., Logan, B.E., 2011. Series assembly of microbialdesalination cells containing stacked electrodialysis cells forpartial or complete seawater desalination. Environ. Sci.Technol. 45, 5840e5845.

Kim, S., Jinschek, J.R., Chen, H., Sholl, D.S., Marand, E., 2007.Scalable fabrication of carbon nanotube/polymernanocomposite membrane for high flux gas transport. NanoLett. 7, 2806e2811.

Kim, S.J., Ko, S.H., Kang, K.H., Han, J., 2010. Direct seawaterdesalination by ion concentration polarization. Nat.Nanotechnol. 5, 297e301.

Kim, Y.D., Thu, K., Ghaffour, N., Ng, K.C., 2013a. Performanceinvestigation of a solar-assisted direct contact membranedistillation system. J. Memb. Sci. 427, 345e364.

Kim, D.Y., Gu, B., Kim, J.H., Yang, D.R., 2013b. Theoretical analysisof a seawater desalination process integrating forwardosmosis, crystallization, and reverse osmosis. J. Memb. Sci.444, 440e448.

Kim, S.G., Hyeon, D.H., Chun, J.H., Chun, B.H., Kim, S.H., 2013c.Novel thin nanocomposite RO membranes for chlorineresistance. Desalination Water Treat. 51, 6338e6345.

Kittur, A.A., Kariduraganavar, M.Y., Toti, U.S., Ramesh, K.,Aminabhavi, T.M., 2003. Pervaporation separation of water-isopropanol mixtures using ZSM-5 zeolite incorporated poly(vinyl alcohol) membranes. J. Appl. Polym. Sci. 90,2441e2448.

Klaysom, C., Cath, T.Y., Depuydt, T., Vankelecom, I.F.J., 2013.Forward and pressure retarded osmosis: potential solutionsfor global challenges in energy and water supply. Chem. Soc.Rev. 42, 6959e6989.

Knox, W.G., Hess, M., Jones, G.E., Smith, H.B., 1961. The clathrateprocess. Chem. Eng. Prog. 57, 66.

Kumar, M., Grzelakowski, M., Zilles, J., Clark, M., Meier, W., 2007.Highly permeable polymeric membranes based on theincorporation of the functional water channel proteinAquaporin Z. Proc. Natl. Acad. Sci. 104, 20719e20724.

Kurth, C.J., Koehler, J.A., Zhou, M., Holmberg, B.A., Burk, R.L.,2014. Hybrid nanoparticle TFC membranes, Patent US 8603340B2. Available at: http://www.google.com/patents/US8603340(last accessed 21.11.14.).

Kuznetsov, Y.P., Kruchinina, E.V., Baklagnia, Y.G.,Khripunov, A.K., Tulupova, O.A., 2007. Russ. J. Appl. Chem. 80,790.

La Mantia, F., Pasta, M., Deshazer, H.D., Logan, B.E., Cui, Y., 2011.Batteries for efficient energy extraction from water salinitydifference. Nano Lett. 11, 1810e1813.

Lee, J., Kim, S., Kim, C., Yoon, Y., 2014. Hybrid capacitivedeionization to enhance the desalination performance ofcapacitive techniques. Energy Environ. Sci. 7, 3683e3689.

Liu, L., Meng, W., Duo, W., Congjie, G., 2010. Current patents offorward osmosis membrane process. Recent Pat. Chem. Eng. 2(1), 76e82.

Loeb, S., Sourirajan, S., 1963. Seawater demineralization bymeans of an osmotic membrane. Adv. Chem. Ser. 38, 117e132.

Luo, H., Xu, P., Roane, T.M., Jenkins, P.E., Ren, Z., 2012. Microbialdesalination cells for improved performance in wastewatertreatment, electricity production and desalination. Bioresour.Technol. 105, 60e66.

Mahmoud, M.S., 2011. Enhancement of solar desalination byhumidification-dehumidification technique. DesalinationWater Treat. 30, 310e318.

Majumder, M., Chopra, N., Andrews, R., Hinds, B.J., 2005.Nanoscale hydrodynamics: enhanced flow in carbonnanotubes. Nature 438, 44.

Mariah, L., Buckley, C.A., Brouckaert, C.J., Curcio, E., Drioli, E.,Jaganyi, D., Ramjugernath, D., 2006. Membrane distillation ofconcentrate brines e role of water activities in the evaluationof driving force. J. Memb. Sci. 280, 937e947.

Mathioulakis, E., Belessiotis, V., Delyannis, E., 2007. Desalinationby using alternative energy: review and state-of-the-art.Desalination 203, 346e365.

McCormack, R.A., Andersen, R.K., June 1995. ClathrateDesalination Plant Preliminary Research Study. U.S.Department of the Interior, Bureau of Reclamation. WaterTreatment Technology Program Report 5.

McCormack, R.A., Niblock, G.A., May 1998. Build and OperateClathrate Desalination Pilot Plant. U.S. Department of theInterior, Bureau of Reclamation. Water Treatment TechnologyProgram Report 31.

wat e r r e s e a r c h 7 5 ( 2 0 1 5 ) 1 6 4e1 8 7186

McCormack, R.A., Niblock, G.A., June 2000. Investigation of HighFreezing Temperature, Zero Ozone, and Zero Global WarmingPotential, Clathrate Formers for Desalination. U.S.Department of the Interior, Bureau of Reclamation. WaterTreatment Technology Program Report 59.

McCutcheon, J., McGinnis, R.L., Elimelech, M., 2005. A novelammonia-carbon dioxide forward (direct) osmosisdesalination process. Desalination 174, 1e11.

McCutcheon, J., McGinnis, R.L., Elimelech, M., 2006. Desalinationby ammonia-carbon dioxide forward osmosis: influence ofdraw and feed solution concentrations on performance onprocess performance. J. Memb. Sci. 278, 114e123.

McGovern, R.K., Lienhard, V., 2014. On the potential of forwardosmosis to energetically outperform reverse osmosisdesalination. J. Memb. Sci. 469, 245e250.

Mehanna, M., Saito, T., Yan, J., Hickner, M., Cai, X., Huang, X.,Logan, B.E., 2010. Using microbial desalination cells to reducewater salinity prior to reverse osmosis. Energy Environ. Sci. 3,1114e1120.

Meindersman, G.W., Guijt, C.M., de Haan, A.B., 2006. Desalinationand water recycling by air gap membrane distillation.Desalination 187, 291e301.

Meng, F., Jiang, J., Zhao, Q., Wang, K., Zhang, J., Fan, Q., Wei, L.,Ding, J., Zheng, Z., 2014. Bioelectrochemical desalination andelectricity generation in microbial desalination cell withdewatered sludge as fuel. Bioresour. Technol. 157, 120e126.

Mi, B., 2014. Graphene oxide membranes for ionic and molecularsieving. Science 343, 740e742.

Miller, James E., Evans, Lindsey R., 2006. Forward Osmosis: a NewApproach to Water Purification and Desalination. SandiaNational Laboratories.

Nair, R.R., Wu, H.A., Jayaram, P.N., Grigorieva, I.V., Geim, A.K.,2012. Unimpeded permeate of water through helium-leak-tight graphene-based membranes. Science 335, 442e443.

NanoH2O, Inc., 2011. Qfx SW 400 ES SWRO Element Data Sheet.Available at: http://www.lg-nanoh2o.com/products/qfx-sw-400-es (last accessed 25.11.14.).

Narayan, G.P., Sharqawy, M.H., Summers, E.K., Lienhard, J.H.,Zubair, S.M., Antar, M.A., 2009. The potential of solar-drivenhumidification-dehumidification desalination for small-scaledecentralized water production. Renew. Sustain. Energy Rev.14 (4), 1187e1201.

Narayan, G.P., Sharqaqy, M.H., Lienhard, J.H., Zubair, S.M., 2010.Thermodynamic analysis of humidification dehumidificationdesalination cycles. Desalination Water Treat. 16, 339e353.

Narayan, G.P., McGovern, R.K., Thiel, G.P., Miller, J.A.,Lienhard, V., 2011. State of humidification dehumidificationdesalination technology. In: Proceedings of the InternationalDesalination Association (IDA) World Congress, Perth,Western Australia, September 4e9.

Ng, K.C., 2014. Adsorption Desalination (AD) Technology. Availableat: https://20e80c98e0ab55004687-1b7b51f623edb7707bfb25abfa60d496.ssl.cf6.rackcdn.com/Medad%20Technologies_TechXchange.pdf (last accessed 02.02.15.).

Ng, K.C., Thu, K., Kim, Y., Chakraborty, A., Amy, G., 2013.Adsorption desalination: an emerging low-cost thermaldesalination method. Desalination 308, 161e179.

Nicolai, A., Sumpter, B.G., Meunier, V., 2014. Tunable waterdesalination across graphene oxide framework membranes.Phys. Chem. Chem. Phys. 16, 8646e8654.

O'Hern, S.C., Boutilier, M.S.H., Idrobo, J.C., Song, Y., Kong, J.,Laoui, T., Atieh, M., Karnik, R., 2014. Selective ionic transportthrough tunable subnanometer pores in single-layer graphenemembranes. Nano Lett. 14, 1234e1241.

Pangarkar, B.L., Sane, M.G., Parjane, S.B., Guddad, M., 2014.Desalination Water Treat. 52, 5199e5218.

Park, K., Hong, S.Y., Lee, J.W., Kang, K.C., Lee, Y.C., Ha, M.,Lee, J.D., 2011. A new apparatus for seawater desalination by

gas hydrates process and removal characteristics of dissolvedminerals (Naþ, Mg2þ, Ca2þ, Kþ, B3þ). Desalination 274, 91e96.

Pasta, M., Wessells, C.D., Cui, Y., La Mantia, F., 2012. Adesalination battery. Nano Lett. 12, 839e843.

Pe~nata, B., Garcıa-Rodrıguez, L., 2012. Current trends and futureprospects in the design of seawater reverse osmosisdesalination technology. Desalination 284, 1e8.

Pendergast, M.M., Hoek, E.M.V., 2011. A review of water treatmentmembrane nanotechnologies. Energy Environ. Sci. 4,1946e1971.

Pendergast, M.M., Ghosh, A.K., Hoek, E.M.V., 2013. Separationperformance and interfacial properties of nanocompositereverse osmosis membranes. Desalination 308, 180e185.

Phuntsho, S., Ho, K.S., Seungkwan, H., Sangyoup, L.,Vigneswaran, S., 2011. A novel low energy fertilizer drivenforward osmosis desalination for direct fertigation: evaluatingthe performance of fertilizer draw solutions. J. Memb. Sci. 375,172e181.

Porada, S., Sales, B.B., Hamelers, H.V.M., Biesheuvel, P.M., 2012.Water desalination with wires. J. Phys. Chem. Lett. 3,1613e1618.

Pryor, T., 2012. Tjuntjunjarra Solar/waste Heat EnergyGroundwater Desalination Project. Available at: http://desalination.edu.au/wp-content/uploads/2011/12/FR2-Pryor-2012.pdf (last accessed 10.02.15.).

Qiu, C., Setiawan, L., Wang, R., Yang, C.Y., Fane, A.G., 2012. Highperformance flat sheet forward osmosis membrane with anNF-like selective layer on a woven fabric embedded substrate.Desalination 287, 266e270.

Qu, D., Wang, J., Wang, L., Hou, D., Luan, Z., Wang, B., 2009.Integration of accelerated precipitation softening withmembrane distillation for high-recovery desalination ofprimary reverse osmosis concentrate. Sep. Purif. Technol. 67,21e25.

Qu, Y., Feng, Y., Wang, X., Liu, J., Lv, J., He, W., Logan, B.E., 2012.Simultaneous water desalination and electricity generation ina microbial desalination cell with electrolyte recirculation forpH control. Bioresour. Technol. 106, 89e94.

Raluy, R.G., Serra, L., Uche, J., 2005. Life cycle assessment ofdesalination technologies integrated with renewable energies.Desalination 183, 81e93.

RWL, Capacitive Deionization, Available at: http://www.rwlwater.com/capacitive-deionization-promising-desalination-technology-resdesigned/, (last accessed 06.09.12.).

Salvetat, J.P., Briggs, A.D., Bonard, J.M., Bacsa, R.R., Kulik, A.J.,Stockli, T., Burnham, N.A., Forro, L., 1999a. Elastic and shearmoduli for single-walled carbon nanotube ropes. Phys. Rev.Lett. 82, 944e947.

Salvetat, J.P., Kulik, A.J., Bonard, J.M., Briggs, G.A.D., Stockli, T.,Metenier, K., Bonnamy, S., Beguin, F., Burnham, N.A., Forro, L.,1999b. Elastic modulus of ordered and disordered mutiwalledcarbon nantubes. Adv. Mater. 11, 161e165.

Seacord, T., MacHarg, J., Coker, S., July 2006. Affordabledesalination Collaboration 2005 results. In: Proceedings of theAmerican Membrane Technology Association Conference inStuart, FL, USA.

Semiat, R., 2008. Energy issues in desalination processes. Environ.Sci. Technol. 42, 8193e8201.

Seo, S.J., Jeon, H., Lee, J.K., Kim, G.Y., Park, D., Nojima, H., Lee, J.,Moon, S.H., 2010. Investigation on removal of hardness ions bycapacitive deionization (CDI) for water softening applications.Water Res. 44, 2267e2275.

Shaffer, D.L., Yip, N.Y., Gilron, J., Elimelech, M., 2012. Perspective:seawater desalination for agriculture by integrated forwardand reverse osmosis: improved product water quality forpotentially less energy. J. Memb. Sci. 415, 1e8.

Sholl, D.S., Johnson, J.K., 2006. Making high-flux membranes withcarbon nanotubes. Science 312 (5776), 1003e1004.

wat e r r e s e a r c h 7 5 ( 2 0 1 5 ) 1 6 4e1 8 7 187

Siemens, Energy-efficient Seawater Desalination, Available at:http://www.siemens.com/press/en/presspicture/?press¼/en/presspicture/pictures-photonews/2009/pn200906/pn200906-02.htm, (last accessed 25.11.14.).

Song, L., Hu, J.Y., One, S.L., Ng, W.J., Elimelech, M., Wilf, M., 2003.Emergence of thermodynamic restriction and its implicationsfor full-scale reverse osmosis processes. Desalination 144,213e228.

Song, L., Schuetze, B., Rainwater, K., December, 2012.Demonstration of a High Recovery and Energy Efficient ROSystem for Small-scale Brackish Water Desalination, FinalReport. Texas Water Development Board.

Stone, M.L., Rae, C., Stewart, F.F., Wilson, A.D., 2013. Switchablepolarity solvents as draw solutes for forward osmosis.Desalination 312, 124e129.

Stover, R., 2007. Seawater reverse osmosis with isobaric energyrecovery devices. Desalination 203, 168e175.

Stover, R., 2013. Permeate recovery and flux maximization insemi-batch reverse osmosis. In: Proceedings of the AmericanWater Works Association (AWWA) and American MembraneTechnology Association (AMTA) Membrane TechnologyConference, February 25e28, San Antonio, Texas.

Stover, R., Efraty, N., 2011. Record low energy consumption withclosed circuit desalination. In: International DesalinationAssociation (IDA) World Congress e Perth Convention andExhibition Center (PCEC), Perth, Western Australia, September4e9, 2011.

Subramani, A., Badruzzaman, M., Oppenheimer, J., Jacangelo, J.G.,2011. Energy minimization strategies and renewable energyutilization for desalination: a review. Water Res. 45,1907e1920.

Subramani, A., Voutchkov, N., Jacangelo, J.G., 2014a. Desalinationenergy minimization using thin film nanocompositemembranes. Desalination 350, 35e43.

Subramani, A., Voutchkov, N., Jacangelo, J.G., 2014b. Seawaterdesalination energy minimization using nanocomposite ROand semi batch RO. In: Proceedings of the AmericanMembrane Technology Association (AMTA) AnnualMembrane Technology Conference, Las Vegas, Nevada.

Sui, H., Han, B.G., Lee, J.K., Walian, P., Jap, B.K., 2001. Structuralbasis of water specific transport through the AQP1 waterchannel. Nature 414, 872e878.

Teoh, M.M., Wang, K., Boyadi, S., Chung, T., Forward Osmosis andMembrane Distillation Processes for Freshwater Production,Available at: http://www.innovationmagazine.com/innovation/volumes/v8n1/coverstory3.shtml (last accessed11.11.14.).

Tomaszewska, M., 2000. Membrane distillation e examples ofapplication in technology and environmental protection. Pol.J. Environ. Stud. 9, 27e36.

Department of Energy (USDOE), Advanced, Energy-efficientHybrid Membrane System for Industrial Water Reuse,Available at: http://energy.gov/sites/prod/files/2013/11/f4/hybrid_membrane_systems_factsheet.pdf, (last accessed12.11.14.).

Voltea, Capacitive Deionization, Available at: http://www.voltea.com/technology/introduction/, (last accessed 06.09.12.).

Walton, J., Lu, H., Turner, C., Solis, S., Hein, H., April 2004. Solarand Waste Heat Desalination by Membrane Distillation.Desalination and Water Purification Research andDevelopment (DRIP) Program, Report No. 81.

Wang, P., Chung, T.S., 2015. Recent advances in membranedistillation processes: membrane development,configuration design and application exploring. J. Memb. Sci.474, 39e56.

Wang, K.Y., Teoh, M.M., Nugroho, A., Chung, T.S., 2011.Integrated forward osmosis-membrane distillation (FO e MD)

hybrid system for the concentration of protein solutions.Chem. Eng. Sci. 66, 2421e2430.

Wang, H., Chung, T.S., Tong, Y.W., Jeyaseelan, K., Armugam, A.,Chen, Z., Hong, M., Meier, W., 2012. Highly permeable andselective pore-spanning biomimetic membrane embeddedwith Aquaporin Z. Small 8, 1185e1190.

Wei, J., Qiu, C., Tang, C.Y., Wang, R., Fane, A.G., 2011. Synthesisand characterization of flat-sheet thin film composite forwardosmosis membranes. J. Memb. Sci. 372, 292e302.

Welgemoed, T.J., 2005. Capacitive Deionization Technology:Development and Evaluation of an Industrial PrototypeSystem. University of Pretoria. Dissertation.

Wilf, M., Bartels, C., 2005. Optimization of seawater RO systemsdesign. Desalination 173, 1e12.

Wilson, A.D., Stewart, F.F., Stone, M.L., 2013. Use of SwitchableSolvents as Forward Osmosis Draw Solutes. Idaho NationalLaboratory. Available at: http://www2.epa.gov/sites/production/files/documents/wilson.pdf (last accessed28.11.14.).

Xie, Z., Hoang, M., Ng, D., Duong, T., Dao, B., Gray, S., 2010. In:Proceedings of IMISTEC10/AMS6 Conference, 22e26 November2010, Sydney, Australia.

Xie, M., Nghiem, L.D., Price, W.E., Elimelech, M., 2013. A forwardosmosis-membrane distillation hybrid process for directsewer mining: system performance and limitations. Environ.Sci. Technol. 47, 13486e13493.

Xue, M., Qiu, H., Gui, W., 2013. Exceptionally fast waterdesalination at complete salt rejection by pristine graphynemonolayers. Nanotechnology 24, 505720.

Yang, Q., Feng, Y., Logan, B.E., 2012. Using cathode spacers tominimize reactor size in air cathode microbial fuel cells.Bioresour. Technol. 110, 273e277.

Yang, H.Y., Han, Z.J., Yu, S.F., Pey, K.L., Ostrikov, K., Karnik, R.,2013. Carbon nanotube membranes with ultrahigh specificabsorption capacity for water desalination and purification.Nat. Commun. 4 (2220), 1e8.

Yangali-Quintanilla, V., Zhenyu, L., Valladeres, R., Li, Q., Amy, G.,2011. Indirect desalination of red sea water with forwardosmosis and low pressure reverse osmosis for water reuse.Desalination 280, 160e166.

Yen, S.K., Mehnas, H.N.F., Su, M., Wang, K.Y., Chung, T.S., 2010.Study of draw solutes using 2-methylimidazole-basedcompounds in forward osmosis. J. Memb. Sci. 364, 242e252.

Yoshikawa, N., Asari, T., Kishi, N., Hayashi, S., Sugai, T.,Shinohara, H., 2008. An efficient fabrication of verticallyaligned carbon nanotubes on flexible aluminum foils bycatalyst-supported chemical vapor deposition.Nanotechnology 19 (24), 245607.

Zhang, F., Brastad, K., He, Z., 2011. Integrating forward osmosisinto microbial fuel cells for wastewater treatment, waterextraction and bioelectricity generation. Environ. Sci. Technol.45, 6690e6696.

Zhao, S., Zou, L., Tang, C.Y., Mulcahy, D., 2012. Recentdevelopments in forward osmosis: opportunities andchallenges. J. Memb. Sci. 396, 1e21.

Zhu, F.Q., Tajkhorshid, E., Schulten,K., 2004.Theoryandsimulationof water permeation in aquaporin-1. Biophys. J. 86, 50e57.

Zhu, A., Christofides, P.D., Cohen, Y., 2009. On ROmembrane andenergycostsandassociated incentives for future enhancementsof membrane permeability. J. Memb. Sci. 344, 1e5.

Zhu, A., Rahardianto, A., Christofides, P.D., Cohen, Y., 2010.Reverse osmosis desalination with high permeabilitymembranes e cost optimization and research needs.Desalination Water Treat. 15, 256e266.

Zwijnenberg, H.J., Koops, G.H., Wessling, M., 2005. Solar drivenmembrane pervaporation for desalination processes. J. Memb.Sci. 250, 235e246.